diff --git a/hidden_states.safetensors b/hidden_states.safetensors index 162879a6b81d61f5db5496b14225880816516131..b955551dba5d6b76f2cd42f61a2f41f0364f3cad 100644 --- a/hidden_states.safetensors +++ b/hidden_states.safetensors @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:d4c3792d3afb47bef55baba76f2050eed270ddaf6a55af4d1b5d5268e071d2c5 +oid sha256:e165fa60388e44734a70f47fa52571c74d7649a1aafcce6f47c7ddf71ee89440 size 1887445592 diff --git a/job_new.json b/job_new.json index ac599cea9dfe616c69df3c7f41e4054340e0ac25..b54ba323c85602d5bbf55fc7277d0466306c3501 100644 --- a/job_new.json +++ b/job_new.json @@ -14,14 +14,14 @@ "rope_alpha": null, "output_measurement": null, "fast_safetensors": false, - "progress": "quant", + "progress": "finished", "status_output": false, "cal_filename": "google/gemma-2-2b-EXL2/cal_data.safetensors", "last_module_idx": 54, "measurement": { "model.layers.0.self_attn": [ { - "accuracy": 0.9515928328037262, + "accuracy": 0.9515531063079834, "total_bits": 30308928, "q_proj": { "group_size": { @@ -85,7 +85,7 @@ } }, { - "accuracy": 0.953981950879097, + "accuracy": 0.9539748653769493, "total_bits": 31455808, "q_proj": { "group_size": { @@ -149,7 +149,7 @@ } }, { - "accuracy": 0.9591670855879784, + "accuracy": 0.9591861926019192, "total_bits": 33412832, "q_proj": { "group_size": { @@ -213,7 +213,7 @@ } }, { - "accuracy": 0.9699935875833035, + "accuracy": 0.9699958208948374, "total_bits": 37983200, "q_proj": { "group_size": { @@ -277,7 +277,7 @@ } }, { - "accuracy": 0.976397393271327, + "accuracy": 0.9763969834893942, "total_bits": 44838176, "q_proj": { "group_size": { @@ -341,7 +341,7 @@ } }, { - "accuracy": 0.9767184294760227, + "accuracy": 0.9767143614590168, "total_bits": 44912768, "q_proj": { "group_size": { @@ -405,7 +405,7 @@ } }, { - "accuracy": 0.986974217928946, + "accuracy": 0.9869659701362252, "total_bits": 57355552, "q_proj": { "group_size": { @@ -457,7 +457,7 @@ } }, { - "accuracy": 0.9872219935059547, + "accuracy": 0.9872188298031688, "total_bits": 57430144, "q_proj": { "group_size": { @@ -509,7 +509,7 @@ } }, { - "accuracy": 0.9881364433094859, + "accuracy": 0.9881322896108031, "total_bits": 57950464, "q_proj": { "group_size": { @@ -561,7 +561,7 @@ } }, { - "accuracy": 0.9888399662449956, + "accuracy": 0.9888321068137884, "total_bits": 58692736, "q_proj": { "group_size": { @@ -613,7 +613,7 @@ } }, { - "accuracy": 0.9880855781957507, + "accuracy": 0.9880868950858712, "total_bits": 59068544, "q_proj": { "group_size": { @@ -677,7 +677,7 @@ } }, { - "accuracy": 0.9889282267540693, + "accuracy": 0.9889282938092947, "total_bits": 59588864, "q_proj": { "group_size": { @@ -741,7 +741,7 @@ } }, { - "accuracy": 0.9899680046364665, + "accuracy": 0.9899678928777575, "total_bits": 61536832, "q_proj": { "group_size": { @@ -802,7 +802,7 @@ } }, { - "accuracy": 0.9907521335408092, + "accuracy": 0.9907488031312823, "total_bits": 62354560, "q_proj": { "group_size": { @@ -863,7 +863,7 @@ } }, { - "accuracy": 0.9944849004969001, + "accuracy": 0.9944888786412776, "total_bits": 75246880, "q_proj": { "group_size": { @@ -924,7 +924,7 @@ } }, { - "accuracy": 0.9953417172655463, + "accuracy": 0.9953421992249787, "total_bits": 76510336, "q_proj": { "group_size": { @@ -985,7 +985,7 @@ } }, { - "accuracy": 0.9965903947595507, + "accuracy": 0.9965901931282133, "total_bits": 85667104, "q_proj": { "group_size": { @@ -1037,7 +1037,7 @@ } }, { - "accuracy": 0.9975632540881634, + "accuracy": 0.9975674538873136, "total_bits": 91722880, "q_proj": { "group_size": { @@ -1089,7 +1089,7 @@ } }, { - "accuracy": 0.9990884908474982, + "accuracy": 0.9990892710047774, "total_bits": 113978656, "q_proj": { "group_size": { @@ -1143,7 +1143,7 @@ ], "model.layers.0.mlp": [ { - "accuracy": 0.8881748169660568, + "accuracy": 0.8882661387324333, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -1195,7 +1195,7 @@ } }, { - "accuracy": 0.8919973969459534, + "accuracy": 0.8919670283794403, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -1247,7 +1247,7 @@ } }, { - "accuracy": 0.9069116935133934, + "accuracy": 0.9069375917315483, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -1296,7 +1296,7 @@ } }, { - "accuracy": 0.9124017059803009, + "accuracy": 0.9124317467212677, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -1345,7 +1345,7 @@ } }, { - "accuracy": 0.9430184848606586, + "accuracy": 0.9430231153964996, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -1397,7 +1397,7 @@ } }, { - "accuracy": 0.9480531290173531, + "accuracy": 0.9480909705162048, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -1449,7 +1449,7 @@ } }, { - "accuracy": 0.9544955678284168, + "accuracy": 0.9545128084719181, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -1498,7 +1498,7 @@ } }, { - "accuracy": 0.968281302601099, + "accuracy": 0.9682874381542206, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -1541,7 +1541,7 @@ } }, { - "accuracy": 0.9713601395487785, + "accuracy": 0.971364825963974, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -1584,7 +1584,7 @@ } }, { - "accuracy": 0.9707059450447559, + "accuracy": 0.9707299135625362, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -1636,7 +1636,7 @@ } }, { - "accuracy": 0.9745050817728043, + "accuracy": 0.974507249891758, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -1688,7 +1688,7 @@ } }, { - "accuracy": 0.9848953913897276, + "accuracy": 0.9848882537335157, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -1740,7 +1740,7 @@ } }, { - "accuracy": 0.9872103426605463, + "accuracy": 0.9872004305943847, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -1792,7 +1792,7 @@ } }, { - "accuracy": 0.9911489551886916, + "accuracy": 0.9911568406969309, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -1835,7 +1835,7 @@ } }, { - "accuracy": 0.9923222842626274, + "accuracy": 0.9923209701664746, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -1884,7 +1884,7 @@ } }, { - "accuracy": 0.9939925689250231, + "accuracy": 0.9939967971295118, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -1930,7 +1930,7 @@ } }, { - "accuracy": 0.9975718185305595, + "accuracy": 0.9975708881393075, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -1972,7 +1972,7 @@ ], "model.layers.1.self_attn": [ { - "accuracy": 0.9072708263993263, + "accuracy": 0.9073722139000893, "total_bits": 30308928, "q_proj": { "group_size": { @@ -2036,7 +2036,7 @@ } }, { - "accuracy": 0.9114431366324425, + "accuracy": 0.91156817227602, "total_bits": 31455808, "q_proj": { "group_size": { @@ -2100,7 +2100,7 @@ } }, { - "accuracy": 0.9211425334215164, + "accuracy": 0.92121671885252, "total_bits": 33412832, "q_proj": { "group_size": { @@ -2164,7 +2164,7 @@ } }, { - "accuracy": 0.940546628087759, + "accuracy": 0.9407354593276978, "total_bits": 37983200, "q_proj": { "group_size": { @@ -2228,7 +2228,7 @@ } }, { - "accuracy": 0.9535201378166676, + "accuracy": 0.9535512514412403, "total_bits": 44838176, "q_proj": { "group_size": { @@ -2292,7 +2292,7 @@ } }, { - "accuracy": 0.9539150819182396, + "accuracy": 0.953955851495266, "total_bits": 44912768, "q_proj": { "group_size": { @@ -2356,7 +2356,7 @@ } }, { - "accuracy": 0.9746329039335251, + "accuracy": 0.9746336918324232, "total_bits": 57355552, "q_proj": { "group_size": { @@ -2408,7 +2408,7 @@ } }, { - "accuracy": 0.9750404637306929, + "accuracy": 0.9750551991164684, "total_bits": 57430144, "q_proj": { "group_size": { @@ -2460,7 +2460,7 @@ } }, { - "accuracy": 0.976738478988409, + "accuracy": 0.976717684417963, "total_bits": 57950464, "q_proj": { "group_size": { @@ -2512,7 +2512,7 @@ } }, { - "accuracy": 0.9780394490808249, + "accuracy": 0.9780480619519949, "total_bits": 58692736, "q_proj": { "group_size": { @@ -2564,7 +2564,7 @@ } }, { - "accuracy": 0.9763826839625835, + "accuracy": 0.9763856939971447, "total_bits": 59068544, "q_proj": { "group_size": { @@ -2628,7 +2628,7 @@ } }, { - "accuracy": 0.9780621360987425, + "accuracy": 0.9780414216220379, "total_bits": 59588864, "q_proj": { "group_size": { @@ -2692,7 +2692,7 @@ } }, { - "accuracy": 0.9801010321825743, + "accuracy": 0.9800688195973635, "total_bits": 61536832, "q_proj": { "group_size": { @@ -2753,7 +2753,7 @@ } }, { - "accuracy": 0.981600085273385, + "accuracy": 0.9815671853721142, "total_bits": 62354560, "q_proj": { "group_size": { @@ -2814,7 +2814,7 @@ } }, { - "accuracy": 0.9890466667711735, + "accuracy": 0.9890200421214104, "total_bits": 75246880, "q_proj": { "group_size": { @@ -2875,7 +2875,7 @@ } }, { - "accuracy": 0.9907460613176227, + "accuracy": 0.9907318344339728, "total_bits": 76510336, "q_proj": { "group_size": { @@ -2936,7 +2936,7 @@ } }, { - "accuracy": 0.9933391944505274, + "accuracy": 0.9933400782756507, "total_bits": 85667104, "q_proj": { "group_size": { @@ -2988,7 +2988,7 @@ } }, { - "accuracy": 0.9952039420604706, + "accuracy": 0.9951940830796957, "total_bits": 91722880, "q_proj": { "group_size": { @@ -3040,7 +3040,7 @@ } }, { - "accuracy": 0.9982604767428711, + "accuracy": 0.9982599728973582, "total_bits": 113978656, "q_proj": { "group_size": { @@ -3094,7 +3094,7 @@ ], "model.layers.1.mlp": [ { - "accuracy": 0.8869373872876167, + "accuracy": 0.887041375041008, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -3146,7 +3146,7 @@ } }, { - "accuracy": 0.891030453145504, + "accuracy": 0.8909792527556419, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -3198,7 +3198,7 @@ } }, { - "accuracy": 0.9027609676122665, + "accuracy": 0.9027382656931877, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -3247,7 +3247,7 @@ } }, { - "accuracy": 0.9062561988830566, + "accuracy": 0.9062387868762016, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -3296,7 +3296,7 @@ } }, { - "accuracy": 0.9437772929668427, + "accuracy": 0.943806029856205, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -3348,7 +3348,7 @@ } }, { - "accuracy": 0.9482904970645905, + "accuracy": 0.9483064226806164, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -3400,7 +3400,7 @@ } }, { - "accuracy": 0.952780719846487, + "accuracy": 0.9527781158685684, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -3449,7 +3449,7 @@ } }, { - "accuracy": 0.9698743838816881, + "accuracy": 0.9698769953101873, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -3492,7 +3492,7 @@ } }, { - "accuracy": 0.9721903912723064, + "accuracy": 0.9722175262868404, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -3535,7 +3535,7 @@ } }, { - "accuracy": 0.9712930209934711, + "accuracy": 0.9713117778301239, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -3587,7 +3587,7 @@ } }, { - "accuracy": 0.9747709520161152, + "accuracy": 0.9747742302715778, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -3639,7 +3639,7 @@ } }, { - "accuracy": 0.9853192111477256, + "accuracy": 0.9853277914226055, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -3691,7 +3691,7 @@ } }, { - "accuracy": 0.987357254140079, + "accuracy": 0.9873669799417257, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -3743,7 +3743,7 @@ } }, { - "accuracy": 0.9915956920012832, + "accuracy": 0.9915947569534183, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -3786,7 +3786,7 @@ } }, { - "accuracy": 0.9925124426372349, + "accuracy": 0.9925175970420241, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -3835,7 +3835,7 @@ } }, { - "accuracy": 0.9935486940667033, + "accuracy": 0.9935514670796692, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -3881,7 +3881,7 @@ } }, { - "accuracy": 0.9976398337166756, + "accuracy": 0.9976405743509531, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -3923,7 +3923,7 @@ ], "model.layers.2.self_attn": [ { - "accuracy": 0.8655510395765305, + "accuracy": 0.8655463457107544, "total_bits": 30308928, "q_proj": { "group_size": { @@ -3987,7 +3987,7 @@ } }, { - "accuracy": 0.8742070347070694, + "accuracy": 0.874164268374443, "total_bits": 31455808, "q_proj": { "group_size": { @@ -4051,7 +4051,7 @@ } }, { - "accuracy": 0.8933860436081886, + "accuracy": 0.8934662640094757, "total_bits": 33412832, "q_proj": { "group_size": { @@ -4115,7 +4115,7 @@ } }, { - "accuracy": 0.9261169582605362, + "accuracy": 0.926240935921669, "total_bits": 37983200, "q_proj": { "group_size": { @@ -4179,7 +4179,7 @@ } }, { - "accuracy": 0.933709405362606, + "accuracy": 0.9337229505181313, "total_bits": 44838176, "q_proj": { "group_size": { @@ -4243,7 +4243,7 @@ } }, { - "accuracy": 0.9347171932458878, + "accuracy": 0.9346923530101776, "total_bits": 44912768, "q_proj": { "group_size": { @@ -4307,7 +4307,7 @@ } }, { - "accuracy": 0.9634591825306416, + "accuracy": 0.9634601436555386, "total_bits": 57355552, "q_proj": { "group_size": { @@ -4359,7 +4359,7 @@ } }, { - "accuracy": 0.9642514251172543, + "accuracy": 0.9642136543989182, "total_bits": 57430144, "q_proj": { "group_size": { @@ -4411,7 +4411,7 @@ } }, { - "accuracy": 0.9666738845407963, + "accuracy": 0.9666857272386551, "total_bits": 57950464, "q_proj": { "group_size": { @@ -4463,7 +4463,7 @@ } }, { - "accuracy": 0.9683135040104389, + "accuracy": 0.9683275371789932, "total_bits": 58692736, "q_proj": { "group_size": { @@ -4515,7 +4515,7 @@ } }, { - "accuracy": 0.966770775616169, + "accuracy": 0.9667858779430389, "total_bits": 59068544, "q_proj": { "group_size": { @@ -4579,7 +4579,7 @@ } }, { - "accuracy": 0.9689657036215067, + "accuracy": 0.9689403381198645, "total_bits": 59588864, "q_proj": { "group_size": { @@ -4643,7 +4643,7 @@ } }, { - "accuracy": 0.9732588063925505, + "accuracy": 0.9732514582574368, "total_bits": 61536832, "q_proj": { "group_size": { @@ -4704,7 +4704,7 @@ } }, { - "accuracy": 0.9749700985848904, + "accuracy": 0.9749796185642481, "total_bits": 62354560, "q_proj": { "group_size": { @@ -4765,7 +4765,7 @@ } }, { - "accuracy": 0.985321925021708, + "accuracy": 0.9853210505098104, "total_bits": 75246880, "q_proj": { "group_size": { @@ -4826,7 +4826,7 @@ } }, { - "accuracy": 0.9874144792556763, + "accuracy": 0.9874111460521817, "total_bits": 76510336, "q_proj": { "group_size": { @@ -4887,7 +4887,7 @@ } }, { - "accuracy": 0.9904179144650698, + "accuracy": 0.9904154865071177, "total_bits": 85667104, "q_proj": { "group_size": { @@ -4939,7 +4939,7 @@ } }, { - "accuracy": 0.9935114970430732, + "accuracy": 0.9935085913166404, "total_bits": 91722880, "q_proj": { "group_size": { @@ -4991,7 +4991,7 @@ } }, { - "accuracy": 0.9975107992067933, + "accuracy": 0.9975118087604642, "total_bits": 113978656, "q_proj": { "group_size": { @@ -5045,7 +5045,7 @@ ], "model.layers.2.mlp": [ { - "accuracy": 0.8512221872806549, + "accuracy": 0.851137638092041, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -5097,7 +5097,7 @@ } }, { - "accuracy": 0.8556944578886032, + "accuracy": 0.8557578176259995, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -5149,7 +5149,7 @@ } }, { - "accuracy": 0.8742920905351639, + "accuracy": 0.8742973953485489, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -5198,7 +5198,7 @@ } }, { - "accuracy": 0.8798726350069046, + "accuracy": 0.8798764571547508, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -5247,7 +5247,7 @@ } }, { - "accuracy": 0.9260013550519943, + "accuracy": 0.9259674102067947, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -5299,7 +5299,7 @@ } }, { - "accuracy": 0.9317035898566246, + "accuracy": 0.9316787645220757, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -5351,7 +5351,7 @@ } }, { - "accuracy": 0.9391488991677761, + "accuracy": 0.9391309916973114, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -5400,7 +5400,7 @@ } }, { - "accuracy": 0.9611527696251869, + "accuracy": 0.9611707739531994, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -5443,7 +5443,7 @@ } }, { - "accuracy": 0.9643270559608936, + "accuracy": 0.964320108294487, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -5486,7 +5486,7 @@ } }, { - "accuracy": 0.9622061587870121, + "accuracy": 0.962238498032093, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -5538,7 +5538,7 @@ } }, { - "accuracy": 0.966707780957222, + "accuracy": 0.9667122215032578, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -5590,7 +5590,7 @@ } }, { - "accuracy": 0.9806768018752337, + "accuracy": 0.9806710071861744, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -5642,7 +5642,7 @@ } }, { - "accuracy": 0.9833411537110806, + "accuracy": 0.9833368640393019, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -5694,7 +5694,7 @@ } }, { - "accuracy": 0.9893169151619077, + "accuracy": 0.98932437133044, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -5737,7 +5737,7 @@ } }, { - "accuracy": 0.9900901559740305, + "accuracy": 0.990092589519918, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -5786,7 +5786,7 @@ } }, { - "accuracy": 0.9917803723365068, + "accuracy": 0.9917868180200458, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -5832,7 +5832,7 @@ } }, { - "accuracy": 0.997135940939188, + "accuracy": 0.9971363425720483, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -5874,7 +5874,7 @@ ], "model.layers.3.self_attn": [ { - "accuracy": 0.8845625817775726, + "accuracy": 0.8846139460802078, "total_bits": 30308928, "q_proj": { "group_size": { @@ -5938,7 +5938,7 @@ } }, { - "accuracy": 0.8901228457689285, + "accuracy": 0.8900518491864204, "total_bits": 31455808, "q_proj": { "group_size": { @@ -6002,7 +6002,7 @@ } }, { - "accuracy": 0.9061683937907219, + "accuracy": 0.9061124995350838, "total_bits": 33412832, "q_proj": { "group_size": { @@ -6066,7 +6066,7 @@ } }, { - "accuracy": 0.9347674697637558, + "accuracy": 0.9347528889775276, "total_bits": 37983200, "q_proj": { "group_size": { @@ -6130,7 +6130,7 @@ } }, { - "accuracy": 0.9422827735543251, + "accuracy": 0.9422792084515095, "total_bits": 44838176, "q_proj": { "group_size": { @@ -6194,7 +6194,7 @@ } }, { - "accuracy": 0.943145889788866, + "accuracy": 0.9431353025138378, "total_bits": 44912768, "q_proj": { "group_size": { @@ -6258,7 +6258,7 @@ } }, { - "accuracy": 0.9687017947435379, + "accuracy": 0.9687346704304218, "total_bits": 57355552, "q_proj": { "group_size": { @@ -6310,7 +6310,7 @@ } }, { - "accuracy": 0.969449121505022, + "accuracy": 0.9694707691669464, "total_bits": 57430144, "q_proj": { "group_size": { @@ -6362,7 +6362,7 @@ } }, { - "accuracy": 0.9714747574180365, + "accuracy": 0.9715053513646126, "total_bits": 57950464, "q_proj": { "group_size": { @@ -6414,7 +6414,7 @@ } }, { - "accuracy": 0.9727979246526957, + "accuracy": 0.9728143084794283, "total_bits": 58692736, "q_proj": { "group_size": { @@ -6466,7 +6466,7 @@ } }, { - "accuracy": 0.9711320530623198, + "accuracy": 0.9711146093904972, "total_bits": 59068544, "q_proj": { "group_size": { @@ -6530,7 +6530,7 @@ } }, { - "accuracy": 0.9730110038071871, + "accuracy": 0.9730188455432653, "total_bits": 59588864, "q_proj": { "group_size": { @@ -6594,7 +6594,7 @@ } }, { - "accuracy": 0.9766805954277515, + "accuracy": 0.9766947887837887, "total_bits": 61536832, "q_proj": { "group_size": { @@ -6655,7 +6655,7 @@ } }, { - "accuracy": 0.9783372972160578, + "accuracy": 0.9783317521214485, "total_bits": 62354560, "q_proj": { "group_size": { @@ -6716,7 +6716,7 @@ } }, { - "accuracy": 0.9873075932264328, + "accuracy": 0.9873018255457282, "total_bits": 75246880, "q_proj": { "group_size": { @@ -6777,7 +6777,7 @@ } }, { - "accuracy": 0.9891520766541362, + "accuracy": 0.9891604781150818, "total_bits": 76510336, "q_proj": { "group_size": { @@ -6838,7 +6838,7 @@ } }, { - "accuracy": 0.9918732857331634, + "accuracy": 0.9918712023645639, "total_bits": 85667104, "q_proj": { "group_size": { @@ -6890,7 +6890,7 @@ } }, { - "accuracy": 0.994609275367111, + "accuracy": 0.9946046634577215, "total_bits": 91722880, "q_proj": { "group_size": { @@ -6942,7 +6942,7 @@ } }, { - "accuracy": 0.9978871110361069, + "accuracy": 0.9978904649615288, "total_bits": 113978656, "q_proj": { "group_size": { @@ -6996,7 +6996,7 @@ ], "model.layers.3.mlp": [ { - "accuracy": 0.8215165734291077, + "accuracy": 0.8215235769748688, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -7048,7 +7048,7 @@ } }, { - "accuracy": 0.8267818093299866, + "accuracy": 0.8267826288938522, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -7100,7 +7100,7 @@ } }, { - "accuracy": 0.8501903116703033, + "accuracy": 0.8501260429620743, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -7149,7 +7149,7 @@ } }, { - "accuracy": 0.8572592884302139, + "accuracy": 0.8572545945644379, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -7198,7 +7198,7 @@ } }, { - "accuracy": 0.9100245088338852, + "accuracy": 0.9100974574685097, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -7250,7 +7250,7 @@ } }, { - "accuracy": 0.9172510206699371, + "accuracy": 0.9171888679265976, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -7302,7 +7302,7 @@ } }, { - "accuracy": 0.9271175712347031, + "accuracy": 0.92706498503685, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -7351,7 +7351,7 @@ } }, { - "accuracy": 0.953081201761961, + "accuracy": 0.9530698768794537, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -7394,7 +7394,7 @@ } }, { - "accuracy": 0.9570418335497379, + "accuracy": 0.957017607986927, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -7437,7 +7437,7 @@ } }, { - "accuracy": 0.9540456458926201, + "accuracy": 0.9540662579238415, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -7489,7 +7489,7 @@ } }, { - "accuracy": 0.9596328996121883, + "accuracy": 0.9596161395311356, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -7541,7 +7541,7 @@ } }, { - "accuracy": 0.9765274990350008, + "accuracy": 0.9765002690255642, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -7593,7 +7593,7 @@ } }, { - "accuracy": 0.9798263423144817, + "accuracy": 0.979815537109971, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -7645,7 +7645,7 @@ } }, { - "accuracy": 0.9871442606672645, + "accuracy": 0.987149802967906, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -7688,7 +7688,7 @@ } }, { - "accuracy": 0.9879334066063166, + "accuracy": 0.9879286577925086, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -7737,7 +7737,7 @@ } }, { - "accuracy": 0.9901348492130637, + "accuracy": 0.9901374354958534, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -7783,7 +7783,7 @@ } }, { - "accuracy": 0.9965592622756958, + "accuracy": 0.9965564699377865, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -7825,7 +7825,7 @@ ], "model.layers.4.self_attn": [ { - "accuracy": 0.8741211891174316, + "accuracy": 0.8741328567266464, "total_bits": 30308928, "q_proj": { "group_size": { @@ -7889,7 +7889,7 @@ } }, { - "accuracy": 0.8816789761185646, + "accuracy": 0.8816472738981247, "total_bits": 31455808, "q_proj": { "group_size": { @@ -7953,7 +7953,7 @@ } }, { - "accuracy": 0.9003885015845299, + "accuracy": 0.900344654917717, "total_bits": 33412832, "q_proj": { "group_size": { @@ -8017,7 +8017,7 @@ } }, { - "accuracy": 0.9307805150747299, + "accuracy": 0.9307383000850677, "total_bits": 37983200, "q_proj": { "group_size": { @@ -8081,7 +8081,7 @@ } }, { - "accuracy": 0.9374072849750519, + "accuracy": 0.9374280199408531, "total_bits": 44838176, "q_proj": { "group_size": { @@ -8145,7 +8145,7 @@ } }, { - "accuracy": 0.9384637139737606, + "accuracy": 0.9384343661367893, "total_bits": 44912768, "q_proj": { "group_size": { @@ -8209,7 +8209,7 @@ } }, { - "accuracy": 0.9653116390109062, + "accuracy": 0.9653166085481644, "total_bits": 57355552, "q_proj": { "group_size": { @@ -8261,7 +8261,7 @@ } }, { - "accuracy": 0.9661814905703068, + "accuracy": 0.9661872051656246, "total_bits": 57430144, "q_proj": { "group_size": { @@ -8313,7 +8313,7 @@ } }, { - "accuracy": 0.9684111662209034, + "accuracy": 0.968420397490263, "total_bits": 57950464, "q_proj": { "group_size": { @@ -8365,7 +8365,7 @@ } }, { - "accuracy": 0.969778697937727, + "accuracy": 0.9697968028485775, "total_bits": 58692736, "q_proj": { "group_size": { @@ -8417,7 +8417,7 @@ } }, { - "accuracy": 0.9687559828162193, + "accuracy": 0.9687640070915222, "total_bits": 59068544, "q_proj": { "group_size": { @@ -8481,7 +8481,7 @@ } }, { - "accuracy": 0.9707713481038809, + "accuracy": 0.9707700200378895, "total_bits": 59588864, "q_proj": { "group_size": { @@ -8545,7 +8545,7 @@ } }, { - "accuracy": 0.975062221288681, + "accuracy": 0.9750581458210945, "total_bits": 61536832, "q_proj": { "group_size": { @@ -8606,7 +8606,7 @@ } }, { - "accuracy": 0.9766068141907454, + "accuracy": 0.9766066540032625, "total_bits": 62354560, "q_proj": { "group_size": { @@ -8667,7 +8667,7 @@ } }, { - "accuracy": 0.986367778852582, + "accuracy": 0.9863643515855074, "total_bits": 75246880, "q_proj": { "group_size": { @@ -8728,7 +8728,7 @@ } }, { - "accuracy": 0.9882961716502905, + "accuracy": 0.9882954657077789, "total_bits": 76510336, "q_proj": { "group_size": { @@ -8789,7 +8789,7 @@ } }, { - "accuracy": 0.9909742707386613, + "accuracy": 0.9909781841561198, "total_bits": 85667104, "q_proj": { "group_size": { @@ -8841,7 +8841,7 @@ } }, { - "accuracy": 0.9939752677455544, + "accuracy": 0.9939774051308632, "total_bits": 91722880, "q_proj": { "group_size": { @@ -8893,7 +8893,7 @@ } }, { - "accuracy": 0.9976646257564425, + "accuracy": 0.9976642981637269, "total_bits": 113978656, "q_proj": { "group_size": { @@ -8947,7 +8947,7 @@ ], "model.layers.4.mlp": [ { - "accuracy": 0.8349805325269699, + "accuracy": 0.8348695486783981, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -8999,7 +8999,7 @@ } }, { - "accuracy": 0.8401748985052109, + "accuracy": 0.8400902897119522, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -9051,7 +9051,7 @@ } }, { - "accuracy": 0.8632630258798599, + "accuracy": 0.863140657544136, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -9100,7 +9100,7 @@ } }, { - "accuracy": 0.8702157586812973, + "accuracy": 0.870071679353714, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -9149,7 +9149,7 @@ } }, { - "accuracy": 0.9168255552649498, + "accuracy": 0.9167942702770233, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -9201,7 +9201,7 @@ } }, { - "accuracy": 0.9239948987960815, + "accuracy": 0.9239982962608337, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -9253,7 +9253,7 @@ } }, { - "accuracy": 0.933478482067585, + "accuracy": 0.9335077553987503, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -9302,7 +9302,7 @@ } }, { - "accuracy": 0.9563411399722099, + "accuracy": 0.9563678838312626, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -9345,7 +9345,7 @@ } }, { - "accuracy": 0.960248950868845, + "accuracy": 0.960248876363039, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -9388,7 +9388,7 @@ } }, { - "accuracy": 0.9574717655777931, + "accuracy": 0.9574565887451172, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -9440,7 +9440,7 @@ } }, { - "accuracy": 0.9629541300237179, + "accuracy": 0.9629608169198036, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -9492,7 +9492,7 @@ } }, { - "accuracy": 0.9782424960285425, + "accuracy": 0.9782421588897705, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -9544,7 +9544,7 @@ } }, { - "accuracy": 0.9814701918512583, + "accuracy": 0.9814739804714918, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -9596,7 +9596,7 @@ } }, { - "accuracy": 0.9879318736493587, + "accuracy": 0.9879294466227293, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -9639,7 +9639,7 @@ } }, { - "accuracy": 0.9888098947703838, + "accuracy": 0.988809896633029, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -9688,7 +9688,7 @@ } }, { - "accuracy": 0.9909165976569057, + "accuracy": 0.9909201338887215, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -9734,7 +9734,7 @@ } }, { - "accuracy": 0.9967805305495858, + "accuracy": 0.9967838707379997, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -9776,7 +9776,7 @@ ], "model.layers.5.self_attn": [ { - "accuracy": 0.8810675218701363, + "accuracy": 0.8811230733990669, "total_bits": 30308928, "q_proj": { "group_size": { @@ -9840,7 +9840,7 @@ } }, { - "accuracy": 0.8886952474713326, + "accuracy": 0.8887623399496078, "total_bits": 31455808, "q_proj": { "group_size": { @@ -9904,7 +9904,7 @@ } }, { - "accuracy": 0.9045190438628197, + "accuracy": 0.904540665447712, "total_bits": 33412832, "q_proj": { "group_size": { @@ -9968,7 +9968,7 @@ } }, { - "accuracy": 0.9352795407176018, + "accuracy": 0.9353184551000595, "total_bits": 37983200, "q_proj": { "group_size": { @@ -10032,7 +10032,7 @@ } }, { - "accuracy": 0.9399208948016167, + "accuracy": 0.9399426616728306, "total_bits": 44838176, "q_proj": { "group_size": { @@ -10096,7 +10096,7 @@ } }, { - "accuracy": 0.9410183131694794, + "accuracy": 0.9409984610974789, "total_bits": 44912768, "q_proj": { "group_size": { @@ -10160,7 +10160,7 @@ } }, { - "accuracy": 0.9663527235388756, + "accuracy": 0.9663707427680492, "total_bits": 57355552, "q_proj": { "group_size": { @@ -10212,7 +10212,7 @@ } }, { - "accuracy": 0.967367872595787, + "accuracy": 0.9673691280186176, "total_bits": 57430144, "q_proj": { "group_size": { @@ -10264,7 +10264,7 @@ } }, { - "accuracy": 0.9696921166032553, + "accuracy": 0.9697138108313084, "total_bits": 57950464, "q_proj": { "group_size": { @@ -10316,7 +10316,7 @@ } }, { - "accuracy": 0.9711299203336239, + "accuracy": 0.9711332526057959, "total_bits": 58692736, "q_proj": { "group_size": { @@ -10368,7 +10368,7 @@ } }, { - "accuracy": 0.9700027704238892, + "accuracy": 0.9700067639350891, "total_bits": 59068544, "q_proj": { "group_size": { @@ -10432,7 +10432,7 @@ } }, { - "accuracy": 0.9720256011933088, + "accuracy": 0.9720334056764841, "total_bits": 59588864, "q_proj": { "group_size": { @@ -10496,7 +10496,7 @@ } }, { - "accuracy": 0.9756313841789961, + "accuracy": 0.9756301864981651, "total_bits": 61536832, "q_proj": { "group_size": { @@ -10557,7 +10557,7 @@ } }, { - "accuracy": 0.9773869439959526, + "accuracy": 0.9774045348167419, "total_bits": 62354560, "q_proj": { "group_size": { @@ -10618,7 +10618,7 @@ } }, { - "accuracy": 0.9866560539230704, + "accuracy": 0.9866563268005848, "total_bits": 75246880, "q_proj": { "group_size": { @@ -10679,7 +10679,7 @@ } }, { - "accuracy": 0.9886289816349745, + "accuracy": 0.9886361388489604, "total_bits": 76510336, "q_proj": { "group_size": { @@ -10740,7 +10740,7 @@ } }, { - "accuracy": 0.9912481410428882, + "accuracy": 0.9912527557462454, "total_bits": 85667104, "q_proj": { "group_size": { @@ -10792,7 +10792,7 @@ } }, { - "accuracy": 0.9942180886864662, + "accuracy": 0.9942179205827415, "total_bits": 91722880, "q_proj": { "group_size": { @@ -10844,7 +10844,7 @@ } }, { - "accuracy": 0.9977244043257087, + "accuracy": 0.9977258909493685, "total_bits": 113978656, "q_proj": { "group_size": { @@ -10898,7 +10898,7 @@ ], "model.layers.5.mlp": [ { - "accuracy": 0.8648017942905426, + "accuracy": 0.8648174107074738, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -10950,7 +10950,7 @@ } }, { - "accuracy": 0.8694759905338287, + "accuracy": 0.8695154935121536, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -11002,7 +11002,7 @@ } }, { - "accuracy": 0.8880036026239395, + "accuracy": 0.888036236166954, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -11051,7 +11051,7 @@ } }, { - "accuracy": 0.8937847763299942, + "accuracy": 0.8938298374414444, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -11100,7 +11100,7 @@ } }, { - "accuracy": 0.9318181350827217, + "accuracy": 0.931794673204422, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -11152,7 +11152,7 @@ } }, { - "accuracy": 0.9377684108912945, + "accuracy": 0.9377873539924622, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -11204,7 +11204,7 @@ } }, { - "accuracy": 0.9455864503979683, + "accuracy": 0.9456103965640068, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -11253,7 +11253,7 @@ } }, { - "accuracy": 0.9638955146074295, + "accuracy": 0.963904220610857, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -11296,7 +11296,7 @@ } }, { - "accuracy": 0.9671880304813385, + "accuracy": 0.9671907536685467, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -11339,7 +11339,7 @@ } }, { - "accuracy": 0.9650383368134499, + "accuracy": 0.9650482423603535, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -11391,7 +11391,7 @@ } }, { - "accuracy": 0.9695848729461432, + "accuracy": 0.9695756994187832, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -11443,7 +11443,7 @@ } }, { - "accuracy": 0.9821004029363394, + "accuracy": 0.982111718505621, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -11495,7 +11495,7 @@ } }, { - "accuracy": 0.9847700102254748, + "accuracy": 0.984773620031774, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -11547,7 +11547,7 @@ } }, { - "accuracy": 0.990053579211235, + "accuracy": 0.9900560518726707, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -11590,7 +11590,7 @@ } }, { - "accuracy": 0.9907908504828811, + "accuracy": 0.990793714299798, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -11639,7 +11639,7 @@ } }, { - "accuracy": 0.992512381169945, + "accuracy": 0.992513523902744, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -11685,7 +11685,7 @@ } }, { - "accuracy": 0.9973284220322967, + "accuracy": 0.997328422497958, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -11727,7 +11727,7 @@ ], "model.layers.6.self_attn": [ { - "accuracy": 0.8890528008341789, + "accuracy": 0.8890272378921509, "total_bits": 30308928, "q_proj": { "group_size": { @@ -11791,7 +11791,7 @@ } }, { - "accuracy": 0.8953105807304382, + "accuracy": 0.8953426331281662, "total_bits": 31455808, "q_proj": { "group_size": { @@ -11855,7 +11855,7 @@ } }, { - "accuracy": 0.9070602431893349, + "accuracy": 0.9071049243211746, "total_bits": 33412832, "q_proj": { "group_size": { @@ -11919,7 +11919,7 @@ } }, { - "accuracy": 0.9330306127667427, + "accuracy": 0.9330921769142151, "total_bits": 37983200, "q_proj": { "group_size": { @@ -11983,7 +11983,7 @@ } }, { - "accuracy": 0.9448112137615681, + "accuracy": 0.9447792358696461, "total_bits": 44838176, "q_proj": { "group_size": { @@ -12047,7 +12047,7 @@ } }, { - "accuracy": 0.9454820677638054, + "accuracy": 0.9455131739377975, "total_bits": 44912768, "q_proj": { "group_size": { @@ -12111,7 +12111,7 @@ } }, { - "accuracy": 0.9695140719413757, + "accuracy": 0.969525508582592, "total_bits": 57355552, "q_proj": { "group_size": { @@ -12163,7 +12163,7 @@ } }, { - "accuracy": 0.9701519533991814, + "accuracy": 0.9701721202582121, "total_bits": 57430144, "q_proj": { "group_size": { @@ -12215,7 +12215,7 @@ } }, { - "accuracy": 0.972233023494482, + "accuracy": 0.9722142405807972, "total_bits": 57950464, "q_proj": { "group_size": { @@ -12267,7 +12267,7 @@ } }, { - "accuracy": 0.9737806580960751, + "accuracy": 0.9737892020493746, "total_bits": 58692736, "q_proj": { "group_size": { @@ -12319,7 +12319,7 @@ } }, { - "accuracy": 0.9722792375832796, + "accuracy": 0.9722646549344063, "total_bits": 59068544, "q_proj": { "group_size": { @@ -12383,7 +12383,7 @@ } }, { - "accuracy": 0.9740959145128727, + "accuracy": 0.9741003829985857, "total_bits": 59588864, "q_proj": { "group_size": { @@ -12447,7 +12447,7 @@ } }, { - "accuracy": 0.9769062623381615, + "accuracy": 0.9769276678562164, "total_bits": 61536832, "q_proj": { "group_size": { @@ -12508,7 +12508,7 @@ } }, { - "accuracy": 0.9786777105182409, + "accuracy": 0.9786724224686623, "total_bits": 62354560, "q_proj": { "group_size": { @@ -12569,7 +12569,7 @@ } }, { - "accuracy": 0.987405676394701, + "accuracy": 0.9874210571870208, "total_bits": 75246880, "q_proj": { "group_size": { @@ -12630,7 +12630,7 @@ } }, { - "accuracy": 0.9893234008923173, + "accuracy": 0.9893340524286032, "total_bits": 76510336, "q_proj": { "group_size": { @@ -12691,7 +12691,7 @@ } }, { - "accuracy": 0.9920955216512084, + "accuracy": 0.9920956883579493, "total_bits": 85667104, "q_proj": { "group_size": { @@ -12743,7 +12743,7 @@ } }, { - "accuracy": 0.9945340054109693, + "accuracy": 0.9945395006798208, "total_bits": 91722880, "q_proj": { "group_size": { @@ -12795,7 +12795,7 @@ } }, { - "accuracy": 0.9979429002851248, + "accuracy": 0.9979436090216041, "total_bits": 113978656, "q_proj": { "group_size": { @@ -12849,7 +12849,7 @@ ], "model.layers.6.mlp": [ { - "accuracy": 0.8601508587598801, + "accuracy": 0.860052615404129, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -12901,7 +12901,7 @@ } }, { - "accuracy": 0.8646412193775177, + "accuracy": 0.864541083574295, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -12953,7 +12953,7 @@ } }, { - "accuracy": 0.883434846997261, + "accuracy": 0.8833191320300102, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -13002,7 +13002,7 @@ } }, { - "accuracy": 0.8892851322889328, + "accuracy": 0.8891885280609131, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -13051,7 +13051,7 @@ } }, { - "accuracy": 0.9288300052285194, + "accuracy": 0.928782157599926, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -13103,7 +13103,7 @@ } }, { - "accuracy": 0.9350205287337303, + "accuracy": 0.9350093528628349, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -13155,7 +13155,7 @@ } }, { - "accuracy": 0.942976824939251, + "accuracy": 0.9429488480091095, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -13204,7 +13204,7 @@ } }, { - "accuracy": 0.9626522436738014, + "accuracy": 0.9626510068774223, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -13247,7 +13247,7 @@ } }, { - "accuracy": 0.9660748802125454, + "accuracy": 0.9660806246101856, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -13290,7 +13290,7 @@ } }, { - "accuracy": 0.9634818024933338, + "accuracy": 0.9634533040225506, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -13342,7 +13342,7 @@ } }, { - "accuracy": 0.9682779163122177, + "accuracy": 0.9682538993656635, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -13394,7 +13394,7 @@ } }, { - "accuracy": 0.9812953136861324, + "accuracy": 0.9812956769019365, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -13446,7 +13446,7 @@ } }, { - "accuracy": 0.9841256625950336, + "accuracy": 0.9841305427253246, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -13498,7 +13498,7 @@ } }, { - "accuracy": 0.9896932039409876, + "accuracy": 0.9896902348846197, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -13541,7 +13541,7 @@ } }, { - "accuracy": 0.9903836958110332, + "accuracy": 0.9903781078755856, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -13590,7 +13590,7 @@ } }, { - "accuracy": 0.9921105708926916, + "accuracy": 0.9921116251498461, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -13636,7 +13636,7 @@ } }, { - "accuracy": 0.9972038897685707, + "accuracy": 0.9972023773007095, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -13678,7 +13678,7 @@ ], "model.layers.7.self_attn": [ { - "accuracy": 0.8790968656539917, + "accuracy": 0.879193864762783, "total_bits": 30308928, "q_proj": { "group_size": { @@ -13742,7 +13742,7 @@ } }, { - "accuracy": 0.8855356946587563, + "accuracy": 0.885476142168045, "total_bits": 31455808, "q_proj": { "group_size": { @@ -13806,7 +13806,7 @@ } }, { - "accuracy": 0.9016124606132507, + "accuracy": 0.9015669599175453, "total_bits": 33412832, "q_proj": { "group_size": { @@ -13870,7 +13870,7 @@ } }, { - "accuracy": 0.9258091002702713, + "accuracy": 0.9257946684956551, "total_bits": 37983200, "q_proj": { "group_size": { @@ -13934,7 +13934,7 @@ } }, { - "accuracy": 0.9374926388263702, + "accuracy": 0.9375413469970226, "total_bits": 44838176, "q_proj": { "group_size": { @@ -13998,7 +13998,7 @@ } }, { - "accuracy": 0.9384367614984512, + "accuracy": 0.9384141005575657, "total_bits": 44912768, "q_proj": { "group_size": { @@ -14062,7 +14062,7 @@ } }, { - "accuracy": 0.9647591635584831, + "accuracy": 0.9647888205945492, "total_bits": 57355552, "q_proj": { "group_size": { @@ -14114,7 +14114,7 @@ } }, { - "accuracy": 0.965646505355835, + "accuracy": 0.9656654335558414, "total_bits": 57430144, "q_proj": { "group_size": { @@ -14166,7 +14166,7 @@ } }, { - "accuracy": 0.9682505577802658, + "accuracy": 0.968209270387888, "total_bits": 57950464, "q_proj": { "group_size": { @@ -14218,7 +14218,7 @@ } }, { - "accuracy": 0.9700952749699354, + "accuracy": 0.9700861666351557, "total_bits": 58692736, "q_proj": { "group_size": { @@ -14270,7 +14270,7 @@ } }, { - "accuracy": 0.9685632511973381, + "accuracy": 0.9685384854674339, "total_bits": 59068544, "q_proj": { "group_size": { @@ -14334,7 +14334,7 @@ } }, { - "accuracy": 0.9708435665816069, + "accuracy": 0.9708608984947205, "total_bits": 59588864, "q_proj": { "group_size": { @@ -14398,7 +14398,7 @@ } }, { - "accuracy": 0.9744916334748268, + "accuracy": 0.9745250958949327, "total_bits": 61536832, "q_proj": { "group_size": { @@ -14459,7 +14459,7 @@ } }, { - "accuracy": 0.976450975984335, + "accuracy": 0.9764064699411392, "total_bits": 62354560, "q_proj": { "group_size": { @@ -14520,7 +14520,7 @@ } }, { - "accuracy": 0.9858461897820234, + "accuracy": 0.9858506005257368, "total_bits": 75246880, "q_proj": { "group_size": { @@ -14581,7 +14581,7 @@ } }, { - "accuracy": 0.9881619503721595, + "accuracy": 0.9881491707637906, "total_bits": 76510336, "q_proj": { "group_size": { @@ -14642,7 +14642,7 @@ } }, { - "accuracy": 0.9907449893653393, + "accuracy": 0.9907346172258258, "total_bits": 85667104, "q_proj": { "group_size": { @@ -14694,7 +14694,7 @@ } }, { - "accuracy": 0.9937627115286887, + "accuracy": 0.9937750529497862, "total_bits": 91722880, "q_proj": { "group_size": { @@ -14746,7 +14746,7 @@ } }, { - "accuracy": 0.9975894596427679, + "accuracy": 0.997590501094237, "total_bits": 113978656, "q_proj": { "group_size": { @@ -14800,7 +14800,7 @@ ], "model.layers.7.mlp": [ { - "accuracy": 0.8853595852851868, + "accuracy": 0.8853648155927658, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -14852,7 +14852,7 @@ } }, { - "accuracy": 0.8890897259116173, + "accuracy": 0.8890001177787781, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -14904,7 +14904,7 @@ } }, { - "accuracy": 0.9030437618494034, + "accuracy": 0.9031427130103111, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -14953,7 +14953,7 @@ } }, { - "accuracy": 0.9074160978198051, + "accuracy": 0.9074975475668907, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -15002,7 +15002,7 @@ } }, { - "accuracy": 0.941648468375206, + "accuracy": 0.9416834153234959, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -15054,7 +15054,7 @@ } }, { - "accuracy": 0.9470459222793579, + "accuracy": 0.9470086395740509, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -15106,7 +15106,7 @@ } }, { - "accuracy": 0.9530564919114113, + "accuracy": 0.9530163891613483, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -15155,7 +15155,7 @@ } }, { - "accuracy": 0.9691261779516935, + "accuracy": 0.9691323079168797, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -15198,7 +15198,7 @@ } }, { - "accuracy": 0.9720455892384052, + "accuracy": 0.9720457717776299, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -15241,7 +15241,7 @@ } }, { - "accuracy": 0.9700704663991928, + "accuracy": 0.9700757917016745, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -15293,7 +15293,7 @@ } }, { - "accuracy": 0.9741590898483992, + "accuracy": 0.9741606414318085, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -15345,7 +15345,7 @@ } }, { - "accuracy": 0.9846790870651603, + "accuracy": 0.9846808835864067, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -15397,7 +15397,7 @@ } }, { - "accuracy": 0.9870709776878357, + "accuracy": 0.9870640141889453, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -15449,7 +15449,7 @@ } }, { - "accuracy": 0.9914329303428531, + "accuracy": 0.9914321266114712, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -15492,7 +15492,7 @@ } }, { - "accuracy": 0.9921443285420537, + "accuracy": 0.9921473953872919, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -15541,7 +15541,7 @@ } }, { - "accuracy": 0.9934511734172702, + "accuracy": 0.9934548917226493, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -15587,7 +15587,7 @@ } }, { - "accuracy": 0.997653994243592, + "accuracy": 0.9976542603690177, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -15629,7 +15629,7 @@ ], "model.layers.8.self_attn": [ { - "accuracy": 0.9262347742915154, + "accuracy": 0.9262180775403976, "total_bits": 30308928, "q_proj": { "group_size": { @@ -15693,7 +15693,7 @@ } }, { - "accuracy": 0.9292053729295731, + "accuracy": 0.9292021617293358, "total_bits": 31455808, "q_proj": { "group_size": { @@ -15757,7 +15757,7 @@ } }, { - "accuracy": 0.9359699636697769, + "accuracy": 0.9360192567110062, "total_bits": 33412832, "q_proj": { "group_size": { @@ -15821,7 +15821,7 @@ } }, { - "accuracy": 0.9522256478667259, + "accuracy": 0.9522725865244865, "total_bits": 37983200, "q_proj": { "group_size": { @@ -15885,7 +15885,7 @@ } }, { - "accuracy": 0.9621881134808064, + "accuracy": 0.9622119329869747, "total_bits": 44838176, "q_proj": { "group_size": { @@ -15949,7 +15949,7 @@ } }, { - "accuracy": 0.9625761918723583, + "accuracy": 0.9625555910170078, "total_bits": 44912768, "q_proj": { "group_size": { @@ -16013,7 +16013,7 @@ } }, { - "accuracy": 0.9790825769305229, + "accuracy": 0.9790912084281445, "total_bits": 57355552, "q_proj": { "group_size": { @@ -16065,7 +16065,7 @@ } }, { - "accuracy": 0.9794511869549751, + "accuracy": 0.9794544521719217, "total_bits": 57430144, "q_proj": { "group_size": { @@ -16117,7 +16117,7 @@ } }, { - "accuracy": 0.9807918332517147, + "accuracy": 0.9808075986802578, "total_bits": 57950464, "q_proj": { "group_size": { @@ -16169,7 +16169,7 @@ } }, { - "accuracy": 0.9819746408611536, + "accuracy": 0.9819814618676901, "total_bits": 58692736, "q_proj": { "group_size": { @@ -16221,7 +16221,7 @@ } }, { - "accuracy": 0.9808401670306921, + "accuracy": 0.980836171656847, "total_bits": 59068544, "q_proj": { "group_size": { @@ -16285,7 +16285,7 @@ } }, { - "accuracy": 0.9820886384695768, + "accuracy": 0.9820813406258821, "total_bits": 59588864, "q_proj": { "group_size": { @@ -16349,7 +16349,7 @@ } }, { - "accuracy": 0.9836654253304005, + "accuracy": 0.9836451895534992, "total_bits": 61536832, "q_proj": { "group_size": { @@ -16410,7 +16410,7 @@ } }, { - "accuracy": 0.984879620373249, + "accuracy": 0.9848749991506338, "total_bits": 62354560, "q_proj": { "group_size": { @@ -16471,7 +16471,7 @@ } }, { - "accuracy": 0.9910424621775746, + "accuracy": 0.9910451974719763, "total_bits": 75246880, "q_proj": { "group_size": { @@ -16532,7 +16532,7 @@ } }, { - "accuracy": 0.9924080655910075, + "accuracy": 0.9923991961404681, "total_bits": 76510336, "q_proj": { "group_size": { @@ -16593,7 +16593,7 @@ } }, { - "accuracy": 0.9945205664262176, + "accuracy": 0.9945177044719458, "total_bits": 85667104, "q_proj": { "group_size": { @@ -16645,7 +16645,7 @@ } }, { - "accuracy": 0.996047873981297, + "accuracy": 0.9960471182130277, "total_bits": 91722880, "q_proj": { "group_size": { @@ -16697,7 +16697,7 @@ } }, { - "accuracy": 0.998559134779498, + "accuracy": 0.9985587185947224, "total_bits": 113978656, "q_proj": { "group_size": { @@ -16751,7 +16751,7 @@ ], "model.layers.8.mlp": [ { - "accuracy": 0.8776565492153168, + "accuracy": 0.8774635121226311, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -16803,7 +16803,7 @@ } }, { - "accuracy": 0.8816645741462708, + "accuracy": 0.8816986083984375, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -16855,7 +16855,7 @@ } }, { - "accuracy": 0.8966159075498581, + "accuracy": 0.8964935094118118, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -16904,7 +16904,7 @@ } }, { - "accuracy": 0.9011181369423866, + "accuracy": 0.9010185077786446, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -16953,7 +16953,7 @@ } }, { - "accuracy": 0.9383117109537125, + "accuracy": 0.9382704459130764, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -17005,7 +17005,7 @@ } }, { - "accuracy": 0.9435076154768467, + "accuracy": 0.943505771458149, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -17057,7 +17057,7 @@ } }, { - "accuracy": 0.94978042319417, + "accuracy": 0.9497714042663574, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -17106,7 +17106,7 @@ } }, { - "accuracy": 0.9677746780216694, + "accuracy": 0.9677646122872829, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -17149,7 +17149,7 @@ } }, { - "accuracy": 0.9705556537955999, + "accuracy": 0.9705599583685398, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -17192,7 +17192,7 @@ } }, { - "accuracy": 0.9684275537729263, + "accuracy": 0.9684217162430286, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -17244,7 +17244,7 @@ } }, { - "accuracy": 0.972480921074748, + "accuracy": 0.9724891372025013, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -17296,7 +17296,7 @@ } }, { - "accuracy": 0.9838621038943529, + "accuracy": 0.9838705994188786, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -17348,7 +17348,7 @@ } }, { - "accuracy": 0.9862669911235571, + "accuracy": 0.9862625077366829, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -17400,7 +17400,7 @@ } }, { - "accuracy": 0.9911140948534012, + "accuracy": 0.9911155067384243, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -17443,7 +17443,7 @@ } }, { - "accuracy": 0.9916985612362623, + "accuracy": 0.9917009193450212, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -17492,7 +17492,7 @@ } }, { - "accuracy": 0.9930045073851943, + "accuracy": 0.9930097297765315, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -17538,7 +17538,7 @@ } }, { - "accuracy": 0.9975975535344332, + "accuracy": 0.9975970382802188, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -17580,7 +17580,7 @@ ], "model.layers.9.self_attn": [ { - "accuracy": 0.9156324192881584, + "accuracy": 0.9154471009969711, "total_bits": 30308928, "q_proj": { "group_size": { @@ -17644,7 +17644,7 @@ } }, { - "accuracy": 0.918098546564579, + "accuracy": 0.9180809780955315, "total_bits": 31455808, "q_proj": { "group_size": { @@ -17708,7 +17708,7 @@ } }, { - "accuracy": 0.9291017279028893, + "accuracy": 0.9292140081524849, "total_bits": 33412832, "q_proj": { "group_size": { @@ -17772,7 +17772,7 @@ } }, { - "accuracy": 0.9474808238446712, + "accuracy": 0.9475394748151302, "total_bits": 37983200, "q_proj": { "group_size": { @@ -17836,7 +17836,7 @@ } }, { - "accuracy": 0.9566572569310665, + "accuracy": 0.9566517025232315, "total_bits": 44838176, "q_proj": { "group_size": { @@ -17900,7 +17900,7 @@ } }, { - "accuracy": 0.9572707749903202, + "accuracy": 0.9572483897209167, "total_bits": 44912768, "q_proj": { "group_size": { @@ -17964,7 +17964,7 @@ } }, { - "accuracy": 0.9765746779739857, + "accuracy": 0.9765565041452646, "total_bits": 57355552, "q_proj": { "group_size": { @@ -18016,7 +18016,7 @@ } }, { - "accuracy": 0.9771516304463148, + "accuracy": 0.9771292954683304, "total_bits": 57430144, "q_proj": { "group_size": { @@ -18068,7 +18068,7 @@ } }, { - "accuracy": 0.9786375071853399, + "accuracy": 0.9786593876779079, "total_bits": 57950464, "q_proj": { "group_size": { @@ -18120,7 +18120,7 @@ } }, { - "accuracy": 0.9796764496713877, + "accuracy": 0.9796735905110836, "total_bits": 58692736, "q_proj": { "group_size": { @@ -18172,7 +18172,7 @@ } }, { - "accuracy": 0.9782940931618214, + "accuracy": 0.9782825838774443, "total_bits": 59068544, "q_proj": { "group_size": { @@ -18236,7 +18236,7 @@ } }, { - "accuracy": 0.9796239528805017, + "accuracy": 0.9796628952026367, "total_bits": 59588864, "q_proj": { "group_size": { @@ -18300,7 +18300,7 @@ } }, { - "accuracy": 0.9821039438247681, + "accuracy": 0.9821524657309055, "total_bits": 61536832, "q_proj": { "group_size": { @@ -18361,7 +18361,7 @@ } }, { - "accuracy": 0.9833755418658257, + "accuracy": 0.9833454601466656, "total_bits": 62354560, "q_proj": { "group_size": { @@ -18422,7 +18422,7 @@ } }, { - "accuracy": 0.9902575109153986, + "accuracy": 0.9902483588084579, "total_bits": 75246880, "q_proj": { "group_size": { @@ -18483,7 +18483,7 @@ } }, { - "accuracy": 0.9916903469711542, + "accuracy": 0.9916817611083388, "total_bits": 76510336, "q_proj": { "group_size": { @@ -18544,7 +18544,7 @@ } }, { - "accuracy": 0.993913528509438, + "accuracy": 0.9939101827330887, "total_bits": 85667104, "q_proj": { "group_size": { @@ -18596,7 +18596,7 @@ } }, { - "accuracy": 0.995829266961664, + "accuracy": 0.99583475664258, "total_bits": 91722880, "q_proj": { "group_size": { @@ -18648,7 +18648,7 @@ } }, { - "accuracy": 0.9984044209122658, + "accuracy": 0.9984046985628083, "total_bits": 113978656, "q_proj": { "group_size": { @@ -18702,7 +18702,7 @@ ], "model.layers.9.mlp": [ { - "accuracy": 0.8979872688651085, + "accuracy": 0.8980022817850113, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -18754,7 +18754,7 @@ } }, { - "accuracy": 0.9010738134384155, + "accuracy": 0.9009037613868713, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -18806,7 +18806,7 @@ } }, { - "accuracy": 0.9138390645384789, + "accuracy": 0.9135458469390869, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -18855,7 +18855,7 @@ } }, { - "accuracy": 0.9180445671081543, + "accuracy": 0.9177292063832283, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -18904,7 +18904,7 @@ } }, { - "accuracy": 0.9479075893759727, + "accuracy": 0.9478593431413174, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -18956,7 +18956,7 @@ } }, { - "accuracy": 0.9523597359657288, + "accuracy": 0.9523407816886902, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -19008,7 +19008,7 @@ } }, { - "accuracy": 0.9581680968403816, + "accuracy": 0.958098616451025, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -19057,7 +19057,7 @@ } }, { - "accuracy": 0.9725048933178186, + "accuracy": 0.9725117534399033, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -19100,7 +19100,7 @@ } }, { - "accuracy": 0.9750282820314169, + "accuracy": 0.975043885409832, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -19143,7 +19143,7 @@ } }, { - "accuracy": 0.9732320122420788, + "accuracy": 0.9732519965618849, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -19195,7 +19195,7 @@ } }, { - "accuracy": 0.9767951015383005, + "accuracy": 0.9767921902239323, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -19247,7 +19247,7 @@ } }, { - "accuracy": 0.9863404175266623, + "accuracy": 0.9863471165299416, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -19299,7 +19299,7 @@ } }, { - "accuracy": 0.9883861737325788, + "accuracy": 0.9883871376514435, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -19351,7 +19351,7 @@ } }, { - "accuracy": 0.992430523969233, + "accuracy": 0.9924324788153172, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -19394,7 +19394,7 @@ } }, { - "accuracy": 0.9929885254241526, + "accuracy": 0.9929874055087566, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -19443,7 +19443,7 @@ } }, { - "accuracy": 0.9942358685657382, + "accuracy": 0.9942356436513364, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -19489,7 +19489,7 @@ } }, { - "accuracy": 0.9979169676080346, + "accuracy": 0.9979170782025903, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -19531,7 +19531,7 @@ ], "model.layers.10.self_attn": [ { - "accuracy": 0.9255667924880981, + "accuracy": 0.9254285618662834, "total_bits": 30308928, "q_proj": { "group_size": { @@ -19595,7 +19595,7 @@ } }, { - "accuracy": 0.9285323470830917, + "accuracy": 0.9285455048084259, "total_bits": 31455808, "q_proj": { "group_size": { @@ -19659,7 +19659,7 @@ } }, { - "accuracy": 0.9377534314990044, + "accuracy": 0.9378277622163296, "total_bits": 33412832, "q_proj": { "group_size": { @@ -19723,7 +19723,7 @@ } }, { - "accuracy": 0.9557530134916306, + "accuracy": 0.9558689743280411, "total_bits": 37983200, "q_proj": { "group_size": { @@ -19787,7 +19787,7 @@ } }, { - "accuracy": 0.9622449390590191, + "accuracy": 0.962271124124527, "total_bits": 44838176, "q_proj": { "group_size": { @@ -19851,7 +19851,7 @@ } }, { - "accuracy": 0.9627878665924072, + "accuracy": 0.9627652131021023, "total_bits": 44912768, "q_proj": { "group_size": { @@ -19915,7 +19915,7 @@ } }, { - "accuracy": 0.9794471673667431, + "accuracy": 0.9794409442692995, "total_bits": 57355552, "q_proj": { "group_size": { @@ -19967,7 +19967,7 @@ } }, { - "accuracy": 0.9799556694924831, + "accuracy": 0.9799547977745533, "total_bits": 57430144, "q_proj": { "group_size": { @@ -20019,7 +20019,7 @@ } }, { - "accuracy": 0.9812588579952717, + "accuracy": 0.9812564067542553, "total_bits": 57950464, "q_proj": { "group_size": { @@ -20071,7 +20071,7 @@ } }, { - "accuracy": 0.982187744230032, + "accuracy": 0.98219870403409, "total_bits": 58692736, "q_proj": { "group_size": { @@ -20123,7 +20123,7 @@ } }, { - "accuracy": 0.9810379017144442, + "accuracy": 0.9810190796852112, "total_bits": 59068544, "q_proj": { "group_size": { @@ -20187,7 +20187,7 @@ } }, { - "accuracy": 0.9822507984936237, + "accuracy": 0.9822470918297768, "total_bits": 59588864, "q_proj": { "group_size": { @@ -20251,7 +20251,7 @@ } }, { - "accuracy": 0.9842813797295094, + "accuracy": 0.9842821806669235, "total_bits": 61536832, "q_proj": { "group_size": { @@ -20312,7 +20312,7 @@ } }, { - "accuracy": 0.9854367896914482, + "accuracy": 0.9854378383606672, "total_bits": 62354560, "q_proj": { "group_size": { @@ -20373,7 +20373,7 @@ } }, { - "accuracy": 0.991425178013742, + "accuracy": 0.9914244255051017, "total_bits": 75246880, "q_proj": { "group_size": { @@ -20434,7 +20434,7 @@ } }, { - "accuracy": 0.9926807903684676, + "accuracy": 0.9926830539479852, "total_bits": 76510336, "q_proj": { "group_size": { @@ -20495,7 +20495,7 @@ } }, { - "accuracy": 0.9946406041271985, + "accuracy": 0.994636666495353, "total_bits": 85667104, "q_proj": { "group_size": { @@ -20547,7 +20547,7 @@ } }, { - "accuracy": 0.9963198550976813, + "accuracy": 0.9963192888535559, "total_bits": 91722880, "q_proj": { "group_size": { @@ -20599,7 +20599,7 @@ } }, { - "accuracy": 0.9985976866446435, + "accuracy": 0.9985979022458196, "total_bits": 113978656, "q_proj": { "group_size": { @@ -20653,7 +20653,7 @@ ], "model.layers.10.mlp": [ { - "accuracy": 0.9072035849094391, + "accuracy": 0.9071538224816322, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -20705,7 +20705,7 @@ } }, { - "accuracy": 0.9100625887513161, + "accuracy": 0.9100388288497925, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -20757,7 +20757,7 @@ } }, { - "accuracy": 0.9221524447202682, + "accuracy": 0.922150731086731, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -20806,7 +20806,7 @@ } }, { - "accuracy": 0.9259821102023125, + "accuracy": 0.9259923249483109, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -20855,7 +20855,7 @@ } }, { - "accuracy": 0.9525629505515099, + "accuracy": 0.9525576233863831, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -20907,7 +20907,7 @@ } }, { - "accuracy": 0.9565679579973221, + "accuracy": 0.9565683789551258, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -20959,7 +20959,7 @@ } }, { - "accuracy": 0.9619633853435516, + "accuracy": 0.9619757160544395, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -21008,7 +21008,7 @@ } }, { - "accuracy": 0.975079670548439, + "accuracy": 0.9750785324722528, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -21051,7 +21051,7 @@ } }, { - "accuracy": 0.9773116856813431, + "accuracy": 0.9773098323494196, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -21094,7 +21094,7 @@ } }, { - "accuracy": 0.9756849519908428, + "accuracy": 0.9756814874708652, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -21146,7 +21146,7 @@ } }, { - "accuracy": 0.9788219351321459, + "accuracy": 0.9788009151816368, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -21198,7 +21198,7 @@ } }, { - "accuracy": 0.9875948084518313, + "accuracy": 0.9875935269519687, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -21250,7 +21250,7 @@ } }, { - "accuracy": 0.9894048757851124, + "accuracy": 0.9894073354080319, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -21302,7 +21302,7 @@ } }, { - "accuracy": 0.993154349271208, + "accuracy": 0.9931564317084849, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -21345,7 +21345,7 @@ } }, { - "accuracy": 0.9936201777309179, + "accuracy": 0.9936129190027714, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -21394,7 +21394,7 @@ } }, { - "accuracy": 0.9947737217880785, + "accuracy": 0.9947650441899896, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -21440,7 +21440,7 @@ } }, { - "accuracy": 0.9981311410665512, + "accuracy": 0.9981314450269565, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -21482,7 +21482,7 @@ ], "model.layers.11.self_attn": [ { - "accuracy": 0.9352168366312981, + "accuracy": 0.9351540505886078, "total_bits": 30308928, "q_proj": { "group_size": { @@ -21546,7 +21546,7 @@ } }, { - "accuracy": 0.9374474063515663, + "accuracy": 0.9374874904751778, "total_bits": 31455808, "q_proj": { "group_size": { @@ -21610,7 +21610,7 @@ } }, { - "accuracy": 0.9444124512374401, + "accuracy": 0.9444890134036541, "total_bits": 33412832, "q_proj": { "group_size": { @@ -21674,7 +21674,7 @@ } }, { - "accuracy": 0.9625572040677071, + "accuracy": 0.9625845514237881, "total_bits": 37983200, "q_proj": { "group_size": { @@ -21738,7 +21738,7 @@ } }, { - "accuracy": 0.9669128507375717, + "accuracy": 0.9669028967618942, "total_bits": 44838176, "q_proj": { "group_size": { @@ -21802,7 +21802,7 @@ } }, { - "accuracy": 0.967328879982233, + "accuracy": 0.9672983437776566, "total_bits": 44912768, "q_proj": { "group_size": { @@ -21866,7 +21866,7 @@ } }, { - "accuracy": 0.9819058496505022, + "accuracy": 0.9818990211933851, "total_bits": 57355552, "q_proj": { "group_size": { @@ -21918,7 +21918,7 @@ } }, { - "accuracy": 0.9823221303522587, + "accuracy": 0.9823174811899662, "total_bits": 57430144, "q_proj": { "group_size": { @@ -21970,7 +21970,7 @@ } }, { - "accuracy": 0.9834885075688362, + "accuracy": 0.9834803491830826, "total_bits": 57950464, "q_proj": { "group_size": { @@ -22022,7 +22022,7 @@ } }, { - "accuracy": 0.9843579828739166, + "accuracy": 0.9843547828495502, "total_bits": 58692736, "q_proj": { "group_size": { @@ -22074,7 +22074,7 @@ } }, { - "accuracy": 0.9833086878061295, + "accuracy": 0.9833086282014847, "total_bits": 59068544, "q_proj": { "group_size": { @@ -22138,7 +22138,7 @@ } }, { - "accuracy": 0.9843683261424303, + "accuracy": 0.9843555800616741, "total_bits": 59588864, "q_proj": { "group_size": { @@ -22202,7 +22202,7 @@ } }, { - "accuracy": 0.9858474927023053, + "accuracy": 0.9858486671000719, "total_bits": 61536832, "q_proj": { "group_size": { @@ -22263,7 +22263,7 @@ } }, { - "accuracy": 0.9868558822199702, + "accuracy": 0.9868420660495758, "total_bits": 62354560, "q_proj": { "group_size": { @@ -22324,7 +22324,7 @@ } }, { - "accuracy": 0.9922798662446439, + "accuracy": 0.9922812515869737, "total_bits": 75246880, "q_proj": { "group_size": { @@ -22385,7 +22385,7 @@ } }, { - "accuracy": 0.9933844502083957, + "accuracy": 0.9933867929503322, "total_bits": 76510336, "q_proj": { "group_size": { @@ -22446,7 +22446,7 @@ } }, { - "accuracy": 0.9952675346285105, + "accuracy": 0.9952673111110926, "total_bits": 85667104, "q_proj": { "group_size": { @@ -22498,7 +22498,7 @@ } }, { - "accuracy": 0.9966305803973228, + "accuracy": 0.9966264350805432, "total_bits": 91722880, "q_proj": { "group_size": { @@ -22550,7 +22550,7 @@ } }, { - "accuracy": 0.9987549100769684, + "accuracy": 0.9987560133449733, "total_bits": 113978656, "q_proj": { "group_size": { @@ -22604,7 +22604,7 @@ ], "model.layers.11.mlp": [ { - "accuracy": 0.922992967069149, + "accuracy": 0.9229258298873901, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -22656,7 +22656,7 @@ } }, { - "accuracy": 0.9253613352775574, + "accuracy": 0.9253562912344933, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -22708,7 +22708,7 @@ } }, { - "accuracy": 0.9349236041307449, + "accuracy": 0.9348956197500229, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -22757,7 +22757,7 @@ } }, { - "accuracy": 0.9380608797073364, + "accuracy": 0.9380255676805973, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -22806,7 +22806,7 @@ } }, { - "accuracy": 0.9607282392680645, + "accuracy": 0.9607758708298206, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -22858,7 +22858,7 @@ } }, { - "accuracy": 0.9641470424830914, + "accuracy": 0.964177817106247, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -22910,7 +22910,7 @@ } }, { - "accuracy": 0.9683838337659836, + "accuracy": 0.9684282056987286, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -22959,7 +22959,7 @@ } }, { - "accuracy": 0.9791794028133154, + "accuracy": 0.9791763704270124, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -23002,7 +23002,7 @@ } }, { - "accuracy": 0.9810750614851713, + "accuracy": 0.9810761008411646, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -23045,7 +23045,7 @@ } }, { - "accuracy": 0.9798438455909491, + "accuracy": 0.9798630569130182, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -23097,7 +23097,7 @@ } }, { - "accuracy": 0.9825253784656525, + "accuracy": 0.9825179129838943, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -23149,7 +23149,7 @@ } }, { - "accuracy": 0.9897021958604455, + "accuracy": 0.9896943140774965, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -23201,7 +23201,7 @@ } }, { - "accuracy": 0.9912520227953792, + "accuracy": 0.9912572083994746, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -23253,7 +23253,7 @@ } }, { - "accuracy": 0.994236150290817, + "accuracy": 0.9942377745173872, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -23296,7 +23296,7 @@ } }, { - "accuracy": 0.9947156864218414, + "accuracy": 0.9947184869088233, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -23345,7 +23345,7 @@ } }, { - "accuracy": 0.9956142762675881, + "accuracy": 0.9956165258772671, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -23391,7 +23391,7 @@ } }, { - "accuracy": 0.9984074216336012, + "accuracy": 0.9984078536508605, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -23433,7 +23433,7 @@ ], "model.layers.12.self_attn": [ { - "accuracy": 0.918254628777504, + "accuracy": 0.9183487147092819, "total_bits": 30308928, "q_proj": { "group_size": { @@ -23497,7 +23497,7 @@ } }, { - "accuracy": 0.9215041920542717, + "accuracy": 0.9214219599962234, "total_bits": 31455808, "q_proj": { "group_size": { @@ -23561,7 +23561,7 @@ } }, { - "accuracy": 0.9291860163211823, + "accuracy": 0.9291634485125542, "total_bits": 33412832, "q_proj": { "group_size": { @@ -23625,7 +23625,7 @@ } }, { - "accuracy": 0.9523625448346138, + "accuracy": 0.9523275718092918, "total_bits": 37983200, "q_proj": { "group_size": { @@ -23689,7 +23689,7 @@ } }, { - "accuracy": 0.9582577794790268, + "accuracy": 0.958238810300827, "total_bits": 44838176, "q_proj": { "group_size": { @@ -23753,7 +23753,7 @@ } }, { - "accuracy": 0.9587131217122078, + "accuracy": 0.9586759731173515, "total_bits": 44912768, "q_proj": { "group_size": { @@ -23817,7 +23817,7 @@ } }, { - "accuracy": 0.9767006933689117, + "accuracy": 0.9766934681683779, "total_bits": 57355552, "q_proj": { "group_size": { @@ -23869,7 +23869,7 @@ } }, { - "accuracy": 0.9771417211741209, + "accuracy": 0.9771546181291342, "total_bits": 57430144, "q_proj": { "group_size": { @@ -23921,7 +23921,7 @@ } }, { - "accuracy": 0.9786731544882059, + "accuracy": 0.978665629401803, "total_bits": 57950464, "q_proj": { "group_size": { @@ -23973,7 +23973,7 @@ } }, { - "accuracy": 0.9797742627561092, + "accuracy": 0.9797797929495573, "total_bits": 58692736, "q_proj": { "group_size": { @@ -24025,7 +24025,7 @@ } }, { - "accuracy": 0.9788827616721392, + "accuracy": 0.9788805264979601, "total_bits": 59068544, "q_proj": { "group_size": { @@ -24089,7 +24089,7 @@ } }, { - "accuracy": 0.980213237926364, + "accuracy": 0.9802060816437006, "total_bits": 59588864, "q_proj": { "group_size": { @@ -24153,7 +24153,7 @@ } }, { - "accuracy": 0.9817759990692139, + "accuracy": 0.9817687887698412, "total_bits": 61536832, "q_proj": { "group_size": { @@ -24214,7 +24214,7 @@ } }, { - "accuracy": 0.9829946663230658, + "accuracy": 0.9829999972134829, "total_bits": 62354560, "q_proj": { "group_size": { @@ -24275,7 +24275,7 @@ } }, { - "accuracy": 0.9900454664602876, + "accuracy": 0.9900399595499039, "total_bits": 75246880, "q_proj": { "group_size": { @@ -24336,7 +24336,7 @@ } }, { - "accuracy": 0.9914399096742272, + "accuracy": 0.9914370570331812, "total_bits": 76510336, "q_proj": { "group_size": { @@ -24397,7 +24397,7 @@ } }, { - "accuracy": 0.9938779231160879, + "accuracy": 0.9938817266374826, "total_bits": 85667104, "q_proj": { "group_size": { @@ -24449,7 +24449,7 @@ } }, { - "accuracy": 0.9955275356769562, + "accuracy": 0.9955261144787073, "total_bits": 91722880, "q_proj": { "group_size": { @@ -24501,7 +24501,7 @@ } }, { - "accuracy": 0.998387377592735, + "accuracy": 0.998388102161698, "total_bits": 113978656, "q_proj": { "group_size": { @@ -24555,7 +24555,7 @@ ], "model.layers.12.mlp": [ { - "accuracy": 0.9067049399018288, + "accuracy": 0.9067919254302979, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -24607,7 +24607,7 @@ } }, { - "accuracy": 0.9094245880842209, + "accuracy": 0.9094521254301071, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -24659,7 +24659,7 @@ } }, { - "accuracy": 0.9205821231007576, + "accuracy": 0.9205865487456322, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -24708,7 +24708,7 @@ } }, { - "accuracy": 0.924173966050148, + "accuracy": 0.9241846203804016, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -24757,7 +24757,7 @@ } }, { - "accuracy": 0.9520841389894485, + "accuracy": 0.9520521014928818, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -24809,7 +24809,7 @@ } }, { - "accuracy": 0.9562458842992783, + "accuracy": 0.9562351442873478, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -24861,7 +24861,7 @@ } }, { - "accuracy": 0.9611620157957077, + "accuracy": 0.961151484400034, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -24910,7 +24910,7 @@ } }, { - "accuracy": 0.9748448915779591, + "accuracy": 0.9748354908078909, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -24953,7 +24953,7 @@ } }, { - "accuracy": 0.9771162606775761, + "accuracy": 0.9771051350980997, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -24996,7 +24996,7 @@ } }, { - "accuracy": 0.975365836173296, + "accuracy": 0.9753743782639503, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -25048,7 +25048,7 @@ } }, { - "accuracy": 0.9786117561161518, + "accuracy": 0.9786114767193794, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -25100,7 +25100,7 @@ } }, { - "accuracy": 0.9873783187940717, + "accuracy": 0.9873899882659316, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -25152,7 +25152,7 @@ } }, { - "accuracy": 0.9893058259040117, + "accuracy": 0.9893016312271357, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -25204,7 +25204,7 @@ } }, { - "accuracy": 0.9930427521467209, + "accuracy": 0.993039789609611, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -25247,7 +25247,7 @@ } }, { - "accuracy": 0.9935374888591468, + "accuracy": 0.9935355954803526, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -25296,7 +25296,7 @@ } }, { - "accuracy": 0.9945738040842116, + "accuracy": 0.9945709388703108, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -25342,7 +25342,7 @@ } }, { - "accuracy": 0.9981157196452841, + "accuracy": 0.9981151465326548, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -25384,7 +25384,7 @@ ], "model.layers.13.self_attn": [ { - "accuracy": 0.914382241666317, + "accuracy": 0.9144919887185097, "total_bits": 30308928, "q_proj": { "group_size": { @@ -25448,7 +25448,7 @@ } }, { - "accuracy": 0.9174628779292107, + "accuracy": 0.9174542501568794, "total_bits": 31455808, "q_proj": { "group_size": { @@ -25512,7 +25512,7 @@ } }, { - "accuracy": 0.9295148551464081, + "accuracy": 0.9295745342969894, "total_bits": 33412832, "q_proj": { "group_size": { @@ -25576,7 +25576,7 @@ } }, { - "accuracy": 0.9491968564689159, + "accuracy": 0.9492183402180672, "total_bits": 37983200, "q_proj": { "group_size": { @@ -25640,7 +25640,7 @@ } }, { - "accuracy": 0.955919798463583, + "accuracy": 0.9559485502541065, "total_bits": 44838176, "q_proj": { "group_size": { @@ -25704,7 +25704,7 @@ } }, { - "accuracy": 0.9567287154495716, + "accuracy": 0.9566921889781952, "total_bits": 44912768, "q_proj": { "group_size": { @@ -25768,7 +25768,7 @@ } }, { - "accuracy": 0.9756081849336624, + "accuracy": 0.9756113328039646, "total_bits": 57355552, "q_proj": { "group_size": { @@ -25820,7 +25820,7 @@ } }, { - "accuracy": 0.9762581083923578, + "accuracy": 0.9762618523091078, "total_bits": 57430144, "q_proj": { "group_size": { @@ -25872,7 +25872,7 @@ } }, { - "accuracy": 0.9783238749951124, + "accuracy": 0.9782921988517046, "total_bits": 57950464, "q_proj": { "group_size": { @@ -25924,7 +25924,7 @@ } }, { - "accuracy": 0.9795089289546013, + "accuracy": 0.9794944487512112, "total_bits": 58692736, "q_proj": { "group_size": { @@ -25976,7 +25976,7 @@ } }, { - "accuracy": 0.9779392145574093, + "accuracy": 0.9778795074671507, "total_bits": 59068544, "q_proj": { "group_size": { @@ -26040,7 +26040,7 @@ } }, { - "accuracy": 0.9795319139957428, + "accuracy": 0.979520695284009, "total_bits": 59588864, "q_proj": { "group_size": { @@ -26104,7 +26104,7 @@ } }, { - "accuracy": 0.9819448851048946, + "accuracy": 0.9819323495030403, "total_bits": 61536832, "q_proj": { "group_size": { @@ -26165,7 +26165,7 @@ } }, { - "accuracy": 0.9835209306329489, + "accuracy": 0.9835347458720207, "total_bits": 62354560, "q_proj": { "group_size": { @@ -26226,7 +26226,7 @@ } }, { - "accuracy": 0.9901209101080894, + "accuracy": 0.9901253506541252, "total_bits": 75246880, "q_proj": { "group_size": { @@ -26287,7 +26287,7 @@ } }, { - "accuracy": 0.9917429555207491, + "accuracy": 0.9917331263422966, "total_bits": 76510336, "q_proj": { "group_size": { @@ -26348,7 +26348,7 @@ } }, { - "accuracy": 0.9937325660139322, + "accuracy": 0.9937320556491613, "total_bits": 85667104, "q_proj": { "group_size": { @@ -26400,7 +26400,7 @@ } }, { - "accuracy": 0.9958382518962026, + "accuracy": 0.9958483558148146, "total_bits": 91722880, "q_proj": { "group_size": { @@ -26452,7 +26452,7 @@ } }, { - "accuracy": 0.9983558729290962, + "accuracy": 0.9983553681522608, "total_bits": 113978656, "q_proj": { "group_size": { @@ -26506,7 +26506,7 @@ ], "model.layers.13.mlp": [ { - "accuracy": 0.9058817103505135, + "accuracy": 0.9057564288377762, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -26558,7 +26558,7 @@ } }, { - "accuracy": 0.9087112993001938, + "accuracy": 0.9086236134171486, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -26610,7 +26610,7 @@ } }, { - "accuracy": 0.9195607155561447, + "accuracy": 0.9196574911475182, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -26659,7 +26659,7 @@ } }, { - "accuracy": 0.923167884349823, + "accuracy": 0.9232490658760071, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -26708,7 +26708,7 @@ } }, { - "accuracy": 0.9515799544751644, + "accuracy": 0.9516337104141712, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -26760,7 +26760,7 @@ } }, { - "accuracy": 0.9560265839099884, + "accuracy": 0.9560323432087898, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -26812,7 +26812,7 @@ } }, { - "accuracy": 0.9607695490121841, + "accuracy": 0.9607681259512901, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -26861,7 +26861,7 @@ } }, { - "accuracy": 0.9745559766888618, + "accuracy": 0.9745499193668365, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -26904,7 +26904,7 @@ } }, { - "accuracy": 0.976823752745986, + "accuracy": 0.9768188558518887, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -26947,7 +26947,7 @@ } }, { - "accuracy": 0.9750555753707886, + "accuracy": 0.9750737082213163, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -26999,7 +26999,7 @@ } }, { - "accuracy": 0.978454452008009, + "accuracy": 0.9784559179097414, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -27051,7 +27051,7 @@ } }, { - "accuracy": 0.9872105978429317, + "accuracy": 0.9872140353545547, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -27103,7 +27103,7 @@ } }, { - "accuracy": 0.9891844298690557, + "accuracy": 0.9891853602603078, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -27155,7 +27155,7 @@ } }, { - "accuracy": 0.992884736508131, + "accuracy": 0.9928796184249222, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -27198,7 +27198,7 @@ } }, { - "accuracy": 0.9934014892205596, + "accuracy": 0.9934069863520563, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -27247,7 +27247,7 @@ } }, { - "accuracy": 0.9944165083579719, + "accuracy": 0.9944231859408319, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -27293,7 +27293,7 @@ } }, { - "accuracy": 0.9980151185300201, + "accuracy": 0.9980159213300794, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -27335,7 +27335,7 @@ ], "model.layers.14.self_attn": [ { - "accuracy": 0.9055267497897148, + "accuracy": 0.9052515029907227, "total_bits": 30308928, "q_proj": { "group_size": { @@ -27399,7 +27399,7 @@ } }, { - "accuracy": 0.9082472920417786, + "accuracy": 0.9081130772829056, "total_bits": 31455808, "q_proj": { "group_size": { @@ -27463,7 +27463,7 @@ } }, { - "accuracy": 0.9182290732860565, + "accuracy": 0.9183298647403717, "total_bits": 33412832, "q_proj": { "group_size": { @@ -27527,7 +27527,7 @@ } }, { - "accuracy": 0.9408689215779305, + "accuracy": 0.9409149326384068, "total_bits": 37983200, "q_proj": { "group_size": { @@ -27591,7 +27591,7 @@ } }, { - "accuracy": 0.9510448761284351, + "accuracy": 0.9511469714343548, "total_bits": 44838176, "q_proj": { "group_size": { @@ -27655,7 +27655,7 @@ } }, { - "accuracy": 0.9517546966671944, + "accuracy": 0.9518323987722397, "total_bits": 44912768, "q_proj": { "group_size": { @@ -27719,7 +27719,7 @@ } }, { - "accuracy": 0.9723473638296127, + "accuracy": 0.9723943509161472, "total_bits": 57355552, "q_proj": { "group_size": { @@ -27771,7 +27771,7 @@ } }, { - "accuracy": 0.9730114750564098, + "accuracy": 0.9731061886996031, "total_bits": 57430144, "q_proj": { "group_size": { @@ -27823,7 +27823,7 @@ } }, { - "accuracy": 0.9746960289776325, + "accuracy": 0.9747657887637615, "total_bits": 57950464, "q_proj": { "group_size": { @@ -27875,7 +27875,7 @@ } }, { - "accuracy": 0.9767959043383598, + "accuracy": 0.9768108148127794, "total_bits": 58692736, "q_proj": { "group_size": { @@ -27927,7 +27927,7 @@ } }, { - "accuracy": 0.9753426536917686, + "accuracy": 0.975377295166254, "total_bits": 59068544, "q_proj": { "group_size": { @@ -27991,7 +27991,7 @@ } }, { - "accuracy": 0.9770860001444817, + "accuracy": 0.9770306386053562, "total_bits": 59588864, "q_proj": { "group_size": { @@ -28055,7 +28055,7 @@ } }, { - "accuracy": 0.979099478572607, + "accuracy": 0.9790472611784935, "total_bits": 61536832, "q_proj": { "group_size": { @@ -28116,7 +28116,7 @@ } }, { - "accuracy": 0.9806705601513386, + "accuracy": 0.9806842133402824, "total_bits": 62354560, "q_proj": { "group_size": { @@ -28177,7 +28177,7 @@ } }, { - "accuracy": 0.9885000661015511, + "accuracy": 0.9885049602016807, "total_bits": 75246880, "q_proj": { "group_size": { @@ -28238,7 +28238,7 @@ } }, { - "accuracy": 0.9902753746137023, + "accuracy": 0.9902975112199783, "total_bits": 76510336, "q_proj": { "group_size": { @@ -28299,7 +28299,7 @@ } }, { - "accuracy": 0.992810903582722, + "accuracy": 0.9927993132732809, "total_bits": 85667104, "q_proj": { "group_size": { @@ -28351,7 +28351,7 @@ } }, { - "accuracy": 0.9949225764721632, + "accuracy": 0.994893008377403, "total_bits": 91722880, "q_proj": { "group_size": { @@ -28403,7 +28403,7 @@ } }, { - "accuracy": 0.9981167989317328, + "accuracy": 0.998116486473009, "total_bits": 113978656, "q_proj": { "group_size": { @@ -28457,7 +28457,7 @@ ], "model.layers.14.mlp": [ { - "accuracy": 0.9059914350509644, + "accuracy": 0.90602907538414, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -28509,7 +28509,7 @@ } }, { - "accuracy": 0.9088080897927284, + "accuracy": 0.9087480306625366, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -28561,7 +28561,7 @@ } }, { - "accuracy": 0.9203511402010918, + "accuracy": 0.9203285947442055, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -28610,7 +28610,7 @@ } }, { - "accuracy": 0.9241925030946732, + "accuracy": 0.9241577237844467, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -28659,7 +28659,7 @@ } }, { - "accuracy": 0.9516731016337872, + "accuracy": 0.951668631285429, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -28711,7 +28711,7 @@ } }, { - "accuracy": 0.9560071490705013, + "accuracy": 0.9559860564768314, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -28763,7 +28763,7 @@ } }, { - "accuracy": 0.9610604681074619, + "accuracy": 0.9610610008239746, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -28812,7 +28812,7 @@ } }, { - "accuracy": 0.9745946303009987, + "accuracy": 0.9746211394667625, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -28855,7 +28855,7 @@ } }, { - "accuracy": 0.9768316633999348, + "accuracy": 0.9768539238721132, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -28898,7 +28898,7 @@ } }, { - "accuracy": 0.9751278180629015, + "accuracy": 0.9751336295157671, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -28950,7 +28950,7 @@ } }, { - "accuracy": 0.9784498494118452, + "accuracy": 0.978433957323432, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -29002,7 +29002,7 @@ } }, { - "accuracy": 0.9872479913756251, + "accuracy": 0.9872386232018471, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -29054,7 +29054,7 @@ } }, { - "accuracy": 0.9891970083117485, + "accuracy": 0.98919727653265, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -29106,7 +29106,7 @@ } }, { - "accuracy": 0.992910641245544, + "accuracy": 0.9929077257402241, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -29149,7 +29149,7 @@ } }, { - "accuracy": 0.9934103316627443, + "accuracy": 0.99340810906142, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -29198,7 +29198,7 @@ } }, { - "accuracy": 0.9944956391118467, + "accuracy": 0.9944933266378939, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -29244,7 +29244,7 @@ } }, { - "accuracy": 0.9980454861652106, + "accuracy": 0.9980455068871379, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -29286,7 +29286,7 @@ ], "model.layers.15.self_attn": [ { - "accuracy": 0.8662237524986267, + "accuracy": 0.8660960495471954, "total_bits": 30308928, "q_proj": { "group_size": { @@ -29350,7 +29350,7 @@ } }, { - "accuracy": 0.8720844835042953, + "accuracy": 0.8721434473991394, "total_bits": 31455808, "q_proj": { "group_size": { @@ -29414,7 +29414,7 @@ } }, { - "accuracy": 0.8906018137931824, + "accuracy": 0.8906832113862038, "total_bits": 33412832, "q_proj": { "group_size": { @@ -29478,7 +29478,7 @@ } }, { - "accuracy": 0.922219954431057, + "accuracy": 0.9223601892590523, "total_bits": 37983200, "q_proj": { "group_size": { @@ -29542,7 +29542,7 @@ } }, { - "accuracy": 0.9315391108393669, + "accuracy": 0.9315524771809578, "total_bits": 44838176, "q_proj": { "group_size": { @@ -29606,7 +29606,7 @@ } }, { - "accuracy": 0.9328188821673393, + "accuracy": 0.9328577220439911, "total_bits": 44912768, "q_proj": { "group_size": { @@ -29670,7 +29670,7 @@ } }, { - "accuracy": 0.9618901945650578, + "accuracy": 0.9618603363633156, "total_bits": 57355552, "q_proj": { "group_size": { @@ -29722,7 +29722,7 @@ } }, { - "accuracy": 0.9632286131381989, + "accuracy": 0.9631414413452148, "total_bits": 57430144, "q_proj": { "group_size": { @@ -29774,7 +29774,7 @@ } }, { - "accuracy": 0.9660533033311367, + "accuracy": 0.9660372957587242, "total_bits": 57950464, "q_proj": { "group_size": { @@ -29826,7 +29826,7 @@ } }, { - "accuracy": 0.9675209484994411, + "accuracy": 0.9675589837133884, "total_bits": 58692736, "q_proj": { "group_size": { @@ -29878,7 +29878,7 @@ } }, { - "accuracy": 0.9656884074211121, + "accuracy": 0.9657478630542755, "total_bits": 59068544, "q_proj": { "group_size": { @@ -29942,7 +29942,7 @@ } }, { - "accuracy": 0.9681769870221615, + "accuracy": 0.9681409150362015, "total_bits": 59588864, "q_proj": { "group_size": { @@ -30006,7 +30006,7 @@ } }, { - "accuracy": 0.9722221530973911, + "accuracy": 0.9722285084426403, "total_bits": 61536832, "q_proj": { "group_size": { @@ -30067,7 +30067,7 @@ } }, { - "accuracy": 0.9743126630783081, + "accuracy": 0.9743023552000523, "total_bits": 62354560, "q_proj": { "group_size": { @@ -30128,7 +30128,7 @@ } }, { - "accuracy": 0.9847090318799019, + "accuracy": 0.984696488827467, "total_bits": 75246880, "q_proj": { "group_size": { @@ -30189,7 +30189,7 @@ } }, { - "accuracy": 0.987136579118669, + "accuracy": 0.9871381670236588, "total_bits": 76510336, "q_proj": { "group_size": { @@ -30250,7 +30250,7 @@ } }, { - "accuracy": 0.9900257075205445, + "accuracy": 0.9900241773575544, "total_bits": 85667104, "q_proj": { "group_size": { @@ -30302,7 +30302,7 @@ } }, { - "accuracy": 0.993544745258987, + "accuracy": 0.9935367675498128, "total_bits": 91722880, "q_proj": { "group_size": { @@ -30354,7 +30354,7 @@ } }, { - "accuracy": 0.9974129700567573, + "accuracy": 0.9974108194001019, "total_bits": 113978656, "q_proj": { "group_size": { @@ -30408,7 +30408,7 @@ ], "model.layers.15.mlp": [ { - "accuracy": 0.888176292181015, + "accuracy": 0.8881220296025276, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -30460,7 +30460,7 @@ } }, { - "accuracy": 0.8917654976248741, + "accuracy": 0.8917921856045723, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -30512,7 +30512,7 @@ } }, { - "accuracy": 0.9060942456126213, + "accuracy": 0.9061857014894485, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -30561,7 +30561,7 @@ } }, { - "accuracy": 0.9107229635119438, + "accuracy": 0.9108146727085114, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -30610,7 +30610,7 @@ } }, { - "accuracy": 0.9432618841528893, + "accuracy": 0.9432695247232914, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -30662,7 +30662,7 @@ } }, { - "accuracy": 0.9481836818158627, + "accuracy": 0.9481980800628662, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -30714,7 +30714,7 @@ } }, { - "accuracy": 0.9543357789516449, + "accuracy": 0.9543447978794575, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -30763,7 +30763,7 @@ } }, { - "accuracy": 0.9701154325157404, + "accuracy": 0.9701192807406187, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -30806,7 +30806,7 @@ } }, { - "accuracy": 0.9727210849523544, + "accuracy": 0.9727198947221041, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -30849,7 +30849,7 @@ } }, { - "accuracy": 0.9709262512624264, + "accuracy": 0.9709332976490259, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -30901,7 +30901,7 @@ } }, { - "accuracy": 0.9747244548052549, + "accuracy": 0.9747237414121628, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -30953,7 +30953,7 @@ } }, { - "accuracy": 0.9851207751780748, + "accuracy": 0.9851209698244929, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -31005,7 +31005,7 @@ } }, { - "accuracy": 0.9873405825346708, + "accuracy": 0.9873364437371492, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -31057,7 +31057,7 @@ } }, { - "accuracy": 0.991690007969737, + "accuracy": 0.991686038672924, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -31100,7 +31100,7 @@ } }, { - "accuracy": 0.9923195065930486, + "accuracy": 0.9923256579786539, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -31149,7 +31149,7 @@ } }, { - "accuracy": 0.9936661920510232, + "accuracy": 0.993671263102442, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -31195,7 +31195,7 @@ } }, { - "accuracy": 0.9976954362355173, + "accuracy": 0.9976960332132876, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -31237,7 +31237,7 @@ ], "model.layers.16.self_attn": [ { - "accuracy": 0.8969288393855095, + "accuracy": 0.8970862925052643, "total_bits": 30308928, "q_proj": { "group_size": { @@ -31301,7 +31301,7 @@ } }, { - "accuracy": 0.9022532030940056, + "accuracy": 0.9023451805114746, "total_bits": 31455808, "q_proj": { "group_size": { @@ -31365,7 +31365,7 @@ } }, { - "accuracy": 0.9167034402489662, + "accuracy": 0.9166429415345192, "total_bits": 33412832, "q_proj": { "group_size": { @@ -31429,7 +31429,7 @@ } }, { - "accuracy": 0.9358327612280846, + "accuracy": 0.9357787296175957, "total_bits": 37983200, "q_proj": { "group_size": { @@ -31493,7 +31493,7 @@ } }, { - "accuracy": 0.9472634494304657, + "accuracy": 0.9472717791795731, "total_bits": 44838176, "q_proj": { "group_size": { @@ -31557,7 +31557,7 @@ } }, { - "accuracy": 0.9484786465764046, + "accuracy": 0.9484795853495598, "total_bits": 44912768, "q_proj": { "group_size": { @@ -31621,7 +31621,7 @@ } }, { - "accuracy": 0.9694637320935726, + "accuracy": 0.9694760199636221, "total_bits": 57355552, "q_proj": { "group_size": { @@ -31673,7 +31673,7 @@ } }, { - "accuracy": 0.9708479661494493, + "accuracy": 0.9708113986998796, "total_bits": 57430144, "q_proj": { "group_size": { @@ -31725,7 +31725,7 @@ } }, { - "accuracy": 0.9735263660550117, + "accuracy": 0.9735017456114292, "total_bits": 57950464, "q_proj": { "group_size": { @@ -31777,7 +31777,7 @@ } }, { - "accuracy": 0.9748429581522942, + "accuracy": 0.9747295752167702, "total_bits": 58692736, "q_proj": { "group_size": { @@ -31829,7 +31829,7 @@ } }, { - "accuracy": 0.9736697655171156, + "accuracy": 0.9736849442124367, "total_bits": 59068544, "q_proj": { "group_size": { @@ -31893,7 +31893,7 @@ } }, { - "accuracy": 0.9757199659943581, + "accuracy": 0.9757067002356052, "total_bits": 59588864, "q_proj": { "group_size": { @@ -31957,7 +31957,7 @@ } }, { - "accuracy": 0.9787998888641596, + "accuracy": 0.9787530265748501, "total_bits": 61536832, "q_proj": { "group_size": { @@ -32018,7 +32018,7 @@ } }, { - "accuracy": 0.9805225525051355, + "accuracy": 0.9805027451366186, "total_bits": 62354560, "q_proj": { "group_size": { @@ -32079,7 +32079,7 @@ } }, { - "accuracy": 0.9882316552102566, + "accuracy": 0.9882538206875324, "total_bits": 75246880, "q_proj": { "group_size": { @@ -32140,7 +32140,7 @@ } }, { - "accuracy": 0.9901310745626688, + "accuracy": 0.9901269851252437, "total_bits": 76510336, "q_proj": { "group_size": { @@ -32201,7 +32201,7 @@ } }, { - "accuracy": 0.9920943574979901, + "accuracy": 0.9920847797766328, "total_bits": 85667104, "q_proj": { "group_size": { @@ -32253,7 +32253,7 @@ } }, { - "accuracy": 0.9948356342501938, + "accuracy": 0.9948330130428076, "total_bits": 91722880, "q_proj": { "group_size": { @@ -32305,7 +32305,7 @@ } }, { - "accuracy": 0.9979338457342237, + "accuracy": 0.9979362795129418, "total_bits": 113978656, "q_proj": { "group_size": { @@ -32359,7 +32359,7 @@ ], "model.layers.16.mlp": [ { - "accuracy": 0.8954387977719307, + "accuracy": 0.8952863588929176, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -32411,7 +32411,7 @@ } }, { - "accuracy": 0.8993488028645515, + "accuracy": 0.8992033004760742, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -32463,7 +32463,7 @@ } }, { - "accuracy": 0.9125201851129532, + "accuracy": 0.9123476222157478, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -32512,7 +32512,7 @@ } }, { - "accuracy": 0.916957214474678, + "accuracy": 0.916797399520874, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -32561,7 +32561,7 @@ } }, { - "accuracy": 0.9469495937228203, + "accuracy": 0.9469642452895641, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -32613,7 +32613,7 @@ } }, { - "accuracy": 0.9515363164246082, + "accuracy": 0.951481893658638, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -32665,7 +32665,7 @@ } }, { - "accuracy": 0.9573016315698624, + "accuracy": 0.9572397917509079, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -32714,7 +32714,7 @@ } }, { - "accuracy": 0.97197007201612, + "accuracy": 0.9719776529818773, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -32757,7 +32757,7 @@ } }, { - "accuracy": 0.9743286669254303, + "accuracy": 0.974342368543148, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -32800,7 +32800,7 @@ } }, { - "accuracy": 0.9728555325418711, + "accuracy": 0.9728449210524559, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -32852,7 +32852,7 @@ } }, { - "accuracy": 0.9763589985668659, + "accuracy": 0.9763544015586376, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -32904,7 +32904,7 @@ } }, { - "accuracy": 0.9861350525170565, + "accuracy": 0.9861237006261945, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -32956,7 +32956,7 @@ } }, { - "accuracy": 0.9881947506219149, + "accuracy": 0.9881899319589138, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -33008,7 +33008,7 @@ } }, { - "accuracy": 0.9922279161401093, + "accuracy": 0.9922269973903894, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -33051,7 +33051,7 @@ } }, { - "accuracy": 0.9928275826387107, + "accuracy": 0.9928280608728528, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -33100,7 +33100,7 @@ } }, { - "accuracy": 0.9940253831446171, + "accuracy": 0.9940254967659712, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -33146,7 +33146,7 @@ } }, { - "accuracy": 0.9978487037587911, + "accuracy": 0.9978484844323248, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -33188,7 +33188,7 @@ ], "model.layers.17.self_attn": [ { - "accuracy": 0.8973334357142448, + "accuracy": 0.897542305290699, "total_bits": 30308928, "q_proj": { "group_size": { @@ -33252,7 +33252,7 @@ } }, { - "accuracy": 0.9017208069562912, + "accuracy": 0.9018189832568169, "total_bits": 31455808, "q_proj": { "group_size": { @@ -33316,7 +33316,7 @@ } }, { - "accuracy": 0.9132213443517685, + "accuracy": 0.9131938815116882, "total_bits": 33412832, "q_proj": { "group_size": { @@ -33380,7 +33380,7 @@ } }, { - "accuracy": 0.938893273472786, + "accuracy": 0.9388857409358025, "total_bits": 37983200, "q_proj": { "group_size": { @@ -33444,7 +33444,7 @@ } }, { - "accuracy": 0.9475571848452091, + "accuracy": 0.9475753307342529, "total_bits": 44838176, "q_proj": { "group_size": { @@ -33508,7 +33508,7 @@ } }, { - "accuracy": 0.9484684765338898, + "accuracy": 0.9485311545431614, "total_bits": 44912768, "q_proj": { "group_size": { @@ -33572,7 +33572,7 @@ } }, { - "accuracy": 0.9707033317536116, + "accuracy": 0.970684751868248, "total_bits": 57355552, "q_proj": { "group_size": { @@ -33624,7 +33624,7 @@ } }, { - "accuracy": 0.9716454949229956, + "accuracy": 0.9716228414326906, "total_bits": 57430144, "q_proj": { "group_size": { @@ -33676,7 +33676,7 @@ } }, { - "accuracy": 0.9738726690411568, + "accuracy": 0.973836112767458, "total_bits": 57950464, "q_proj": { "group_size": { @@ -33728,7 +33728,7 @@ } }, { - "accuracy": 0.9751554299145937, + "accuracy": 0.9751142673194408, "total_bits": 58692736, "q_proj": { "group_size": { @@ -33780,7 +33780,7 @@ } }, { - "accuracy": 0.973699688911438, + "accuracy": 0.9736771062016487, "total_bits": 59068544, "q_proj": { "group_size": { @@ -33844,7 +33844,7 @@ } }, { - "accuracy": 0.9755098223686218, + "accuracy": 0.9755203202366829, "total_bits": 59588864, "q_proj": { "group_size": { @@ -33908,7 +33908,7 @@ } }, { - "accuracy": 0.9779850598424673, + "accuracy": 0.9779762793332338, "total_bits": 61536832, "q_proj": { "group_size": { @@ -33969,7 +33969,7 @@ } }, { - "accuracy": 0.9795938245952129, + "accuracy": 0.9795962031930685, "total_bits": 62354560, "q_proj": { "group_size": { @@ -34030,7 +34030,7 @@ } }, { - "accuracy": 0.9879280971363187, + "accuracy": 0.987911774776876, "total_bits": 75246880, "q_proj": { "group_size": { @@ -34091,7 +34091,7 @@ } }, { - "accuracy": 0.989773684181273, + "accuracy": 0.9897674126550555, "total_bits": 76510336, "q_proj": { "group_size": { @@ -34152,7 +34152,7 @@ } }, { - "accuracy": 0.9923741007223725, + "accuracy": 0.9923535925336182, "total_bits": 85667104, "q_proj": { "group_size": { @@ -34204,7 +34204,7 @@ } }, { - "accuracy": 0.9947892669588327, + "accuracy": 0.9947894560173154, "total_bits": 91722880, "q_proj": { "group_size": { @@ -34256,7 +34256,7 @@ } }, { - "accuracy": 0.99802085920237, + "accuracy": 0.998017291072756, "total_bits": 113978656, "q_proj": { "group_size": { @@ -34310,7 +34310,7 @@ ], "model.layers.17.mlp": [ { - "accuracy": 0.8820443674921989, + "accuracy": 0.8820276632905006, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -34362,7 +34362,7 @@ } }, { - "accuracy": 0.8866309821605682, + "accuracy": 0.8865934386849403, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -34414,7 +34414,7 @@ } }, { - "accuracy": 0.9031416177749634, + "accuracy": 0.9032439365983009, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -34463,7 +34463,7 @@ } }, { - "accuracy": 0.9086097851395607, + "accuracy": 0.9086964055895805, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -34512,7 +34512,7 @@ } }, { - "accuracy": 0.9407571367919445, + "accuracy": 0.9407499842345715, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -34564,7 +34564,7 @@ } }, { - "accuracy": 0.9456419833004475, + "accuracy": 0.9456696957349777, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -34616,7 +34616,7 @@ } }, { - "accuracy": 0.9527646824717522, + "accuracy": 0.952761884778738, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -34665,7 +34665,7 @@ } }, { - "accuracy": 0.9684986434876919, + "accuracy": 0.9684955701231956, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -34708,7 +34708,7 @@ } }, { - "accuracy": 0.9712209962308407, + "accuracy": 0.9712157864123583, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -34751,7 +34751,7 @@ } }, { - "accuracy": 0.969695445150137, + "accuracy": 0.969698378816247, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -34803,7 +34803,7 @@ } }, { - "accuracy": 0.973537739366293, + "accuracy": 0.9735384881496429, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -34855,7 +34855,7 @@ } }, { - "accuracy": 0.984518950805068, + "accuracy": 0.9845216348767281, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -34907,7 +34907,7 @@ } }, { - "accuracy": 0.9867853112518787, + "accuracy": 0.9867835436016321, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -34959,7 +34959,7 @@ } }, { - "accuracy": 0.991300848312676, + "accuracy": 0.9913051715120673, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -35002,7 +35002,7 @@ } }, { - "accuracy": 0.9920124607160687, + "accuracy": 0.9920066902413964, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -35051,7 +35051,7 @@ } }, { - "accuracy": 0.9935338972136378, + "accuracy": 0.9935281220823526, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -35097,7 +35097,7 @@ } }, { - "accuracy": 0.9976069342810661, + "accuracy": 0.9976056986488402, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -35139,7 +35139,7 @@ ], "model.layers.18.self_attn": [ { - "accuracy": 0.8961049765348434, + "accuracy": 0.895967960357666, "total_bits": 30308928, "q_proj": { "group_size": { @@ -35203,7 +35203,7 @@ } }, { - "accuracy": 0.9031495451927185, + "accuracy": 0.9031614512205124, "total_bits": 31455808, "q_proj": { "group_size": { @@ -35267,7 +35267,7 @@ } }, { - "accuracy": 0.915522888302803, + "accuracy": 0.9155553206801414, "total_bits": 33412832, "q_proj": { "group_size": { @@ -35331,7 +35331,7 @@ } }, { - "accuracy": 0.9396003894507885, + "accuracy": 0.9396803267300129, "total_bits": 37983200, "q_proj": { "group_size": { @@ -35395,7 +35395,7 @@ } }, { - "accuracy": 0.9481151774525642, + "accuracy": 0.9481409452855587, "total_bits": 44838176, "q_proj": { "group_size": { @@ -35459,7 +35459,7 @@ } }, { - "accuracy": 0.9490873701870441, + "accuracy": 0.9491141587495804, "total_bits": 44912768, "q_proj": { "group_size": { @@ -35523,7 +35523,7 @@ } }, { - "accuracy": 0.9702784214168787, + "accuracy": 0.9702505953609943, "total_bits": 57355552, "q_proj": { "group_size": { @@ -35575,7 +35575,7 @@ } }, { - "accuracy": 0.9712849836796522, + "accuracy": 0.9712676033377647, "total_bits": 57430144, "q_proj": { "group_size": { @@ -35627,7 +35627,7 @@ } }, { - "accuracy": 0.9733225014060736, + "accuracy": 0.9733318723738194, "total_bits": 57950464, "q_proj": { "group_size": { @@ -35679,7 +35679,7 @@ } }, { - "accuracy": 0.9746621809899807, + "accuracy": 0.9746661577373743, "total_bits": 58692736, "q_proj": { "group_size": { @@ -35731,7 +35731,7 @@ } }, { - "accuracy": 0.9740584716200829, + "accuracy": 0.9740777555853128, "total_bits": 59068544, "q_proj": { "group_size": { @@ -35795,7 +35795,7 @@ } }, { - "accuracy": 0.9757704772055149, + "accuracy": 0.975718442350626, "total_bits": 59588864, "q_proj": { "group_size": { @@ -35859,7 +35859,7 @@ } }, { - "accuracy": 0.9784729983657598, + "accuracy": 0.9784317370504141, "total_bits": 61536832, "q_proj": { "group_size": { @@ -35920,7 +35920,7 @@ } }, { - "accuracy": 0.9800199028104544, + "accuracy": 0.9800386726856232, "total_bits": 62354560, "q_proj": { "group_size": { @@ -35981,7 +35981,7 @@ } }, { - "accuracy": 0.9881230108439922, + "accuracy": 0.988131619989872, "total_bits": 75246880, "q_proj": { "group_size": { @@ -36042,7 +36042,7 @@ } }, { - "accuracy": 0.9899277659133077, + "accuracy": 0.9899393441155553, "total_bits": 76510336, "q_proj": { "group_size": { @@ -36103,7 +36103,7 @@ } }, { - "accuracy": 0.9922054400667548, + "accuracy": 0.9921957161277533, "total_bits": 85667104, "q_proj": { "group_size": { @@ -36155,7 +36155,7 @@ } }, { - "accuracy": 0.9948229189030826, + "accuracy": 0.994826287496835, "total_bits": 91722880, "q_proj": { "group_size": { @@ -36207,7 +36207,7 @@ } }, { - "accuracy": 0.9979508460965008, + "accuracy": 0.9979522456414998, "total_bits": 113978656, "q_proj": { "group_size": { @@ -36261,7 +36261,7 @@ ], "model.layers.18.mlp": [ { - "accuracy": 0.8908465802669525, + "accuracy": 0.890760600566864, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -36313,7 +36313,7 @@ } }, { - "accuracy": 0.8948191478848457, + "accuracy": 0.8948594033718109, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -36365,7 +36365,7 @@ } }, { - "accuracy": 0.9099691584706306, + "accuracy": 0.9100265651941299, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -36414,7 +36414,7 @@ } }, { - "accuracy": 0.9150631278753281, + "accuracy": 0.915120542049408, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -36463,7 +36463,7 @@ } }, { - "accuracy": 0.9450795240700245, + "accuracy": 0.9450446665287018, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -36515,7 +36515,7 @@ } }, { - "accuracy": 0.949484571814537, + "accuracy": 0.9494681656360626, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -36567,7 +36567,7 @@ } }, { - "accuracy": 0.9560906924307346, + "accuracy": 0.9560775049030781, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -36616,7 +36616,7 @@ } }, { - "accuracy": 0.9707168936729431, + "accuracy": 0.9707209877669811, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -36659,7 +36659,7 @@ } }, { - "accuracy": 0.9732743687927723, + "accuracy": 0.9732760544866323, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -36702,7 +36702,7 @@ } }, { - "accuracy": 0.9719167854636908, + "accuracy": 0.9719307515770197, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -36754,7 +36754,7 @@ } }, { - "accuracy": 0.9754060637205839, + "accuracy": 0.9753952287137508, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -36806,7 +36806,7 @@ } }, { - "accuracy": 0.9856863981112838, + "accuracy": 0.9856886547058821, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -36858,7 +36858,7 @@ } }, { - "accuracy": 0.9877114491537213, + "accuracy": 0.9877165118232369, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -36910,7 +36910,7 @@ } }, { - "accuracy": 0.9919731765985489, + "accuracy": 0.9919724566861987, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -36953,7 +36953,7 @@ } }, { - "accuracy": 0.9926252705045044, + "accuracy": 0.9926214762963355, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -37002,7 +37002,7 @@ } }, { - "accuracy": 0.9940348407253623, + "accuracy": 0.9940323028713465, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -37048,7 +37048,7 @@ } }, { - "accuracy": 0.9977936870418489, + "accuracy": 0.997793102869764, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -37090,7 +37090,7 @@ ], "model.layers.19.self_attn": [ { - "accuracy": 0.9066938832402229, + "accuracy": 0.9067771062254906, "total_bits": 30308928, "q_proj": { "group_size": { @@ -37154,7 +37154,7 @@ } }, { - "accuracy": 0.9125282913446426, + "accuracy": 0.9125330448150635, "total_bits": 31455808, "q_proj": { "group_size": { @@ -37218,7 +37218,7 @@ } }, { - "accuracy": 0.9224199429154396, + "accuracy": 0.9224216490983963, "total_bits": 33412832, "q_proj": { "group_size": { @@ -37282,7 +37282,7 @@ } }, { - "accuracy": 0.942340325564146, + "accuracy": 0.9422799944877625, "total_bits": 37983200, "q_proj": { "group_size": { @@ -37346,7 +37346,7 @@ } }, { - "accuracy": 0.9528491497039795, + "accuracy": 0.9527867995202541, "total_bits": 44838176, "q_proj": { "group_size": { @@ -37410,7 +37410,7 @@ } }, { - "accuracy": 0.9537361785769463, + "accuracy": 0.9536703340709209, "total_bits": 44912768, "q_proj": { "group_size": { @@ -37474,7 +37474,7 @@ } }, { - "accuracy": 0.9727440439164639, + "accuracy": 0.9727084040641785, "total_bits": 57355552, "q_proj": { "group_size": { @@ -37526,7 +37526,7 @@ } }, { - "accuracy": 0.9736681282520294, + "accuracy": 0.9736226126551628, "total_bits": 57430144, "q_proj": { "group_size": { @@ -37578,7 +37578,7 @@ } }, { - "accuracy": 0.9758165907114744, + "accuracy": 0.9758408833295107, "total_bits": 57950464, "q_proj": { "group_size": { @@ -37630,7 +37630,7 @@ } }, { - "accuracy": 0.9769553057849407, + "accuracy": 0.976946609094739, "total_bits": 58692736, "q_proj": { "group_size": { @@ -37682,7 +37682,7 @@ } }, { - "accuracy": 0.9764603525400162, + "accuracy": 0.9764624573290348, "total_bits": 59068544, "q_proj": { "group_size": { @@ -37746,7 +37746,7 @@ } }, { - "accuracy": 0.9781377725303173, + "accuracy": 0.9780647847801447, "total_bits": 59588864, "q_proj": { "group_size": { @@ -37810,7 +37810,7 @@ } }, { - "accuracy": 0.9803321305662394, + "accuracy": 0.9802794177085161, "total_bits": 61536832, "q_proj": { "group_size": { @@ -37871,7 +37871,7 @@ } }, { - "accuracy": 0.9818277545273304, + "accuracy": 0.9817991387099028, "total_bits": 62354560, "q_proj": { "group_size": { @@ -37932,7 +37932,7 @@ } }, { - "accuracy": 0.9891065321862698, + "accuracy": 0.9891098402440548, "total_bits": 75246880, "q_proj": { "group_size": { @@ -37993,7 +37993,7 @@ } }, { - "accuracy": 0.9908470399677753, + "accuracy": 0.9908483978360891, "total_bits": 76510336, "q_proj": { "group_size": { @@ -38054,7 +38054,7 @@ } }, { - "accuracy": 0.9929368873126805, + "accuracy": 0.9929364663548768, "total_bits": 85667104, "q_proj": { "group_size": { @@ -38106,7 +38106,7 @@ } }, { - "accuracy": 0.9951926521025598, + "accuracy": 0.9951782925054431, "total_bits": 91722880, "q_proj": { "group_size": { @@ -38158,7 +38158,7 @@ } }, { - "accuracy": 0.9981394049245864, + "accuracy": 0.9981401044642553, "total_bits": 113978656, "q_proj": { "group_size": { @@ -38212,7 +38212,7 @@ ], "model.layers.19.mlp": [ { - "accuracy": 0.898843489587307, + "accuracy": 0.8988735228776932, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -38264,7 +38264,7 @@ } }, { - "accuracy": 0.9021168202161789, + "accuracy": 0.9020884037017822, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -38316,7 +38316,7 @@ } }, { - "accuracy": 0.9158534109592438, + "accuracy": 0.9158180356025696, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -38365,7 +38365,7 @@ } }, { - "accuracy": 0.9202790409326553, + "accuracy": 0.9202406033873558, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -38414,7 +38414,7 @@ } }, { - "accuracy": 0.9486076608300209, + "accuracy": 0.9486297108232975, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -38466,7 +38466,7 @@ } }, { - "accuracy": 0.9527660310268402, + "accuracy": 0.9527538195252419, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -38518,7 +38518,7 @@ } }, { - "accuracy": 0.9588682539761066, + "accuracy": 0.9588394649326801, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -38567,7 +38567,7 @@ } }, { - "accuracy": 0.9728159811347723, + "accuracy": 0.9728188179433346, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -38610,7 +38610,7 @@ } }, { - "accuracy": 0.9751972481608391, + "accuracy": 0.9751956835389137, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -38653,7 +38653,7 @@ } }, { - "accuracy": 0.9737109690904617, + "accuracy": 0.9737076684832573, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -38705,7 +38705,7 @@ } }, { - "accuracy": 0.9769691210240126, + "accuracy": 0.976981945335865, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -38757,7 +38757,7 @@ } }, { - "accuracy": 0.986573307774961, + "accuracy": 0.9865698497742414, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -38809,7 +38809,7 @@ } }, { - "accuracy": 0.9884986458346248, + "accuracy": 0.9884996181353927, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -38861,7 +38861,7 @@ } }, { - "accuracy": 0.9925676062703133, + "accuracy": 0.9925639107823372, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -38904,7 +38904,7 @@ } }, { - "accuracy": 0.993094963952899, + "accuracy": 0.993094386998564, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -38953,7 +38953,7 @@ } }, { - "accuracy": 0.9943557516671717, + "accuracy": 0.9943572375923395, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -38999,7 +38999,7 @@ } }, { - "accuracy": 0.9979684855788946, + "accuracy": 0.9979692660272121, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -39041,7 +39041,7 @@ ], "model.layers.20.self_attn": [ { - "accuracy": 0.9254191517829895, + "accuracy": 0.9252974018454552, "total_bits": 30308928, "q_proj": { "group_size": { @@ -39105,7 +39105,7 @@ } }, { - "accuracy": 0.9304513037204742, + "accuracy": 0.9304210692644119, "total_bits": 31455808, "q_proj": { "group_size": { @@ -39169,7 +39169,7 @@ } }, { - "accuracy": 0.937650989741087, + "accuracy": 0.9376390874385834, "total_bits": 33412832, "q_proj": { "group_size": { @@ -39233,7 +39233,7 @@ } }, { - "accuracy": 0.9561277441680431, + "accuracy": 0.956094067543745, "total_bits": 37983200, "q_proj": { "group_size": { @@ -39297,7 +39297,7 @@ } }, { - "accuracy": 0.9629123508930206, + "accuracy": 0.9629477187991142, "total_bits": 44838176, "q_proj": { "group_size": { @@ -39361,7 +39361,7 @@ } }, { - "accuracy": 0.9635394588112831, + "accuracy": 0.9635471068322659, "total_bits": 44912768, "q_proj": { "group_size": { @@ -39425,7 +39425,7 @@ } }, { - "accuracy": 0.9784403182566166, + "accuracy": 0.9784409739077091, "total_bits": 57355552, "q_proj": { "group_size": { @@ -39477,7 +39477,7 @@ } }, { - "accuracy": 0.9791186898946762, + "accuracy": 0.9791207388043404, "total_bits": 57430144, "q_proj": { "group_size": { @@ -39529,7 +39529,7 @@ } }, { - "accuracy": 0.9805207420140505, + "accuracy": 0.9805529490113258, "total_bits": 57950464, "q_proj": { "group_size": { @@ -39581,7 +39581,7 @@ } }, { - "accuracy": 0.981576981022954, + "accuracy": 0.9815756008028984, "total_bits": 58692736, "q_proj": { "group_size": { @@ -39633,7 +39633,7 @@ } }, { - "accuracy": 0.9814564771950245, + "accuracy": 0.981438597664237, "total_bits": 59068544, "q_proj": { "group_size": { @@ -39697,7 +39697,7 @@ } }, { - "accuracy": 0.9826366528868675, + "accuracy": 0.9826178383082151, "total_bits": 59588864, "q_proj": { "group_size": { @@ -39761,7 +39761,7 @@ } }, { - "accuracy": 0.9840204436331987, + "accuracy": 0.9840060472488403, "total_bits": 61536832, "q_proj": { "group_size": { @@ -39822,7 +39822,7 @@ } }, { - "accuracy": 0.9851696165278554, + "accuracy": 0.9851967319846153, "total_bits": 62354560, "q_proj": { "group_size": { @@ -39883,7 +39883,7 @@ } }, { - "accuracy": 0.9912118604406714, + "accuracy": 0.99119614623487, "total_bits": 75246880, "q_proj": { "group_size": { @@ -39944,7 +39944,7 @@ } }, { - "accuracy": 0.9925604872405529, + "accuracy": 0.992552753072232, "total_bits": 76510336, "q_proj": { "group_size": { @@ -40005,7 +40005,7 @@ } }, { - "accuracy": 0.9943439392372966, + "accuracy": 0.9943421361967921, "total_bits": 85667104, "q_proj": { "group_size": { @@ -40057,7 +40057,7 @@ } }, { - "accuracy": 0.9961527525447309, + "accuracy": 0.996142350602895, "total_bits": 91722880, "q_proj": { "group_size": { @@ -40109,7 +40109,7 @@ } }, { - "accuracy": 0.9985169863793999, + "accuracy": 0.9985177288763225, "total_bits": 113978656, "q_proj": { "group_size": { @@ -40163,7 +40163,7 @@ ], "model.layers.20.mlp": [ { - "accuracy": 0.9078842997550964, + "accuracy": 0.9078511148691177, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -40215,7 +40215,7 @@ } }, { - "accuracy": 0.9108857437968254, + "accuracy": 0.9108638241887093, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -40267,7 +40267,7 @@ } }, { - "accuracy": 0.9233747720718384, + "accuracy": 0.9233528077602386, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -40316,7 +40316,7 @@ } }, { - "accuracy": 0.927286222577095, + "accuracy": 0.9272778257727623, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -40365,7 +40365,7 @@ } }, { - "accuracy": 0.9533279724419117, + "accuracy": 0.9533239230513573, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -40417,7 +40417,7 @@ } }, { - "accuracy": 0.9570191614329815, + "accuracy": 0.9570236355066299, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -40469,7 +40469,7 @@ } }, { - "accuracy": 0.9625181145966053, + "accuracy": 0.962521106004715, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -40518,7 +40518,7 @@ } }, { - "accuracy": 0.9753356482833624, + "accuracy": 0.9753350578248501, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -40561,7 +40561,7 @@ } }, { - "accuracy": 0.977478351444006, + "accuracy": 0.9774747285991907, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -40604,7 +40604,7 @@ } }, { - "accuracy": 0.9761522505432367, + "accuracy": 0.9761437401175499, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -40656,7 +40656,7 @@ } }, { - "accuracy": 0.9790751449763775, + "accuracy": 0.9790640287101269, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -40708,7 +40708,7 @@ } }, { - "accuracy": 0.9878394166007638, + "accuracy": 0.9878420624881983, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -40760,7 +40760,7 @@ } }, { - "accuracy": 0.9895561756566167, + "accuracy": 0.9895521271973848, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -40812,7 +40812,7 @@ } }, { - "accuracy": 0.9932784507982433, + "accuracy": 0.993278375826776, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -40855,7 +40855,7 @@ } }, { - "accuracy": 0.9937723437324166, + "accuracy": 0.9937705183401704, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -40904,7 +40904,7 @@ } }, { - "accuracy": 0.9949068506248295, + "accuracy": 0.9949048236012459, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -40950,7 +40950,7 @@ } }, { - "accuracy": 0.9981734347529709, + "accuracy": 0.9981726757250726, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -40992,7 +40992,7 @@ ], "model.layers.21.self_attn": [ { - "accuracy": 0.9116751402616501, + "accuracy": 0.9116267189383507, "total_bits": 30308928, "q_proj": { "group_size": { @@ -41056,7 +41056,7 @@ } }, { - "accuracy": 0.9161897599697113, + "accuracy": 0.9161302745342255, "total_bits": 31455808, "q_proj": { "group_size": { @@ -41120,7 +41120,7 @@ } }, { - "accuracy": 0.9299630001187325, + "accuracy": 0.9298892468214035, "total_bits": 33412832, "q_proj": { "group_size": { @@ -41184,7 +41184,7 @@ } }, { - "accuracy": 0.951121561229229, + "accuracy": 0.951112475246191, "total_bits": 37983200, "q_proj": { "group_size": { @@ -41248,7 +41248,7 @@ } }, { - "accuracy": 0.9554766491055489, + "accuracy": 0.9555105529725552, "total_bits": 44838176, "q_proj": { "group_size": { @@ -41312,7 +41312,7 @@ } }, { - "accuracy": 0.9563397578895092, + "accuracy": 0.9563395045697689, "total_bits": 44912768, "q_proj": { "group_size": { @@ -41376,7 +41376,7 @@ } }, { - "accuracy": 0.9752995986491442, + "accuracy": 0.9753066599369049, "total_bits": 57355552, "q_proj": { "group_size": { @@ -41428,7 +41428,7 @@ } }, { - "accuracy": 0.976159306243062, + "accuracy": 0.9761861730366945, "total_bits": 57430144, "q_proj": { "group_size": { @@ -41480,7 +41480,7 @@ } }, { - "accuracy": 0.9777697827666998, + "accuracy": 0.9777861218899488, "total_bits": 57950464, "q_proj": { "group_size": { @@ -41532,7 +41532,7 @@ } }, { - "accuracy": 0.9786687269806862, + "accuracy": 0.9786933194845915, "total_bits": 58692736, "q_proj": { "group_size": { @@ -41584,7 +41584,7 @@ } }, { - "accuracy": 0.9778354410082102, + "accuracy": 0.9778481628745794, "total_bits": 59068544, "q_proj": { "group_size": { @@ -41648,7 +41648,7 @@ } }, { - "accuracy": 0.9793091975152493, + "accuracy": 0.9793014619499445, "total_bits": 59588864, "q_proj": { "group_size": { @@ -41712,7 +41712,7 @@ } }, { - "accuracy": 0.9823502618819475, + "accuracy": 0.9823220800608397, "total_bits": 61536832, "q_proj": { "group_size": { @@ -41773,7 +41773,7 @@ } }, { - "accuracy": 0.9835798889398575, + "accuracy": 0.9835791904479265, "total_bits": 62354560, "q_proj": { "group_size": { @@ -41834,7 +41834,7 @@ } }, { - "accuracy": 0.9903135802596807, + "accuracy": 0.9903201386332512, "total_bits": 75246880, "q_proj": { "group_size": { @@ -41895,7 +41895,7 @@ } }, { - "accuracy": 0.9917551334947348, + "accuracy": 0.9917648062109947, "total_bits": 76510336, "q_proj": { "group_size": { @@ -41956,7 +41956,7 @@ } }, { - "accuracy": 0.993569896556437, + "accuracy": 0.9935718071646988, "total_bits": 85667104, "q_proj": { "group_size": { @@ -42008,7 +42008,7 @@ } }, { - "accuracy": 0.9958970271982253, + "accuracy": 0.995899414177984, "total_bits": 91722880, "q_proj": { "group_size": { @@ -42060,7 +42060,7 @@ } }, { - "accuracy": 0.9983291744720191, + "accuracy": 0.9983296216232702, "total_bits": 113978656, "q_proj": { "group_size": { @@ -42114,7 +42114,7 @@ ], "model.layers.21.mlp": [ { - "accuracy": 0.9132195189595222, + "accuracy": 0.9132231697440147, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -42166,7 +42166,7 @@ } }, { - "accuracy": 0.9158970862627029, + "accuracy": 0.9158701747655869, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -42218,7 +42218,7 @@ } }, { - "accuracy": 0.9271287769079208, + "accuracy": 0.9270192459225655, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -42267,7 +42267,7 @@ } }, { - "accuracy": 0.9307196065783501, + "accuracy": 0.9306086674332619, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -42316,7 +42316,7 @@ } }, { - "accuracy": 0.955941203981638, + "accuracy": 0.9559643566608429, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -42368,7 +42368,7 @@ } }, { - "accuracy": 0.9593969620764256, + "accuracy": 0.9593847468495369, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -42420,7 +42420,7 @@ } }, { - "accuracy": 0.9644332267343998, + "accuracy": 0.9644359201192856, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -42469,7 +42469,7 @@ } }, { - "accuracy": 0.9767589289695024, + "accuracy": 0.9767424874007702, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -42512,7 +42512,7 @@ } }, { - "accuracy": 0.9787649121135473, + "accuracy": 0.9787469636648893, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -42555,7 +42555,7 @@ } }, { - "accuracy": 0.9774891547858715, + "accuracy": 0.9774995297193527, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -42607,7 +42607,7 @@ } }, { - "accuracy": 0.9802252352237701, + "accuracy": 0.9802324343472719, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -42659,7 +42659,7 @@ } }, { - "accuracy": 0.9885327704250813, + "accuracy": 0.9885260602459311, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -42711,7 +42711,7 @@ } }, { - "accuracy": 0.9901311350986362, + "accuracy": 0.9901291783899069, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -42763,7 +42763,7 @@ } }, { - "accuracy": 0.9936433634720743, + "accuracy": 0.9936425597406924, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -42806,7 +42806,7 @@ } }, { - "accuracy": 0.9941183044575155, + "accuracy": 0.9941207375377417, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -42855,7 +42855,7 @@ } }, { - "accuracy": 0.995133682154119, + "accuracy": 0.9951349799521267, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -42901,7 +42901,7 @@ } }, { - "accuracy": 0.9982258901000023, + "accuracy": 0.9982257036026567, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -42943,7 +42943,7 @@ ], "model.layers.22.self_attn": [ { - "accuracy": 0.8600220680236816, + "accuracy": 0.8604224324226379, "total_bits": 30308928, "q_proj": { "group_size": { @@ -43007,7 +43007,7 @@ } }, { - "accuracy": 0.8723062872886658, + "accuracy": 0.8720817565917969, "total_bits": 31455808, "q_proj": { "group_size": { @@ -43071,7 +43071,7 @@ } }, { - "accuracy": 0.8845950290560722, + "accuracy": 0.8842579200863838, "total_bits": 33412832, "q_proj": { "group_size": { @@ -43135,7 +43135,7 @@ } }, { - "accuracy": 0.9110095426440239, + "accuracy": 0.9107557758688927, "total_bits": 37983200, "q_proj": { "group_size": { @@ -43199,7 +43199,7 @@ } }, { - "accuracy": 0.9318739250302315, + "accuracy": 0.9317922592163086, "total_bits": 44838176, "q_proj": { "group_size": { @@ -43263,7 +43263,7 @@ } }, { - "accuracy": 0.9331164434552193, + "accuracy": 0.933014988899231, "total_bits": 44912768, "q_proj": { "group_size": { @@ -43327,7 +43327,7 @@ } }, { - "accuracy": 0.9576672576367855, + "accuracy": 0.9576846696436405, "total_bits": 57355552, "q_proj": { "group_size": { @@ -43379,7 +43379,7 @@ } }, { - "accuracy": 0.9591643884778023, + "accuracy": 0.9591707922518253, "total_bits": 57430144, "q_proj": { "group_size": { @@ -43431,7 +43431,7 @@ } }, { - "accuracy": 0.9620198979973793, + "accuracy": 0.9620674923062325, "total_bits": 57950464, "q_proj": { "group_size": { @@ -43483,7 +43483,7 @@ } }, { - "accuracy": 0.9642005227506161, + "accuracy": 0.9641453549265862, "total_bits": 58692736, "q_proj": { "group_size": { @@ -43535,7 +43535,7 @@ } }, { - "accuracy": 0.9660037606954575, + "accuracy": 0.9659963771700859, "total_bits": 59068544, "q_proj": { "group_size": { @@ -43599,7 +43599,7 @@ } }, { - "accuracy": 0.968350138515234, + "accuracy": 0.9683596417307854, "total_bits": 59588864, "q_proj": { "group_size": { @@ -43663,7 +43663,7 @@ } }, { - "accuracy": 0.9704683814197779, + "accuracy": 0.970446715131402, "total_bits": 61536832, "q_proj": { "group_size": { @@ -43724,7 +43724,7 @@ } }, { - "accuracy": 0.9727780558168888, + "accuracy": 0.9728066977113485, "total_bits": 62354560, "q_proj": { "group_size": { @@ -43785,7 +43785,7 @@ } }, { - "accuracy": 0.9837654158473015, + "accuracy": 0.9837965164333582, "total_bits": 75246880, "q_proj": { "group_size": { @@ -43846,7 +43846,7 @@ } }, { - "accuracy": 0.9864024138078094, + "accuracy": 0.9864012431353331, "total_bits": 76510336, "q_proj": { "group_size": { @@ -43907,7 +43907,7 @@ } }, { - "accuracy": 0.988879844546318, + "accuracy": 0.9888947391882539, "total_bits": 85667104, "q_proj": { "group_size": { @@ -43959,7 +43959,7 @@ } }, { - "accuracy": 0.9924624958075583, + "accuracy": 0.9924732316285372, "total_bits": 91722880, "q_proj": { "group_size": { @@ -44011,7 +44011,7 @@ } }, { - "accuracy": 0.9970627035945654, + "accuracy": 0.997059682616964, "total_bits": 113978656, "q_proj": { "group_size": { @@ -44065,7 +44065,7 @@ ], "model.layers.22.mlp": [ { - "accuracy": 0.9158392548561096, + "accuracy": 0.9158530607819557, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -44117,7 +44117,7 @@ } }, { - "accuracy": 0.9184657260775566, + "accuracy": 0.9185132682323456, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -44169,7 +44169,7 @@ } }, { - "accuracy": 0.9284176975488663, + "accuracy": 0.9284360781311989, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -44218,7 +44218,7 @@ } }, { - "accuracy": 0.9315355196595192, + "accuracy": 0.9315476194024086, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -44267,7 +44267,7 @@ } }, { - "accuracy": 0.9572760388255119, + "accuracy": 0.9572666324675083, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -44319,7 +44319,7 @@ } }, { - "accuracy": 0.9605343677103519, + "accuracy": 0.9605594612658024, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -44371,7 +44371,7 @@ } }, { - "accuracy": 0.9649710729718208, + "accuracy": 0.9650028869509697, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -44420,7 +44420,7 @@ } }, { - "accuracy": 0.9774835146963596, + "accuracy": 0.9774914421141148, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -44463,7 +44463,7 @@ } }, { - "accuracy": 0.9793780688196421, + "accuracy": 0.9793844223022461, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -44506,7 +44506,7 @@ } }, { - "accuracy": 0.9781624637544155, + "accuracy": 0.9781443998217583, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -44558,7 +44558,7 @@ } }, { - "accuracy": 0.9807386361062527, + "accuracy": 0.9807199090719223, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -44610,7 +44610,7 @@ } }, { - "accuracy": 0.988841949030757, + "accuracy": 0.9888424472883344, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -44662,7 +44662,7 @@ } }, { - "accuracy": 0.9903620686382055, + "accuracy": 0.9903717981651425, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -44714,7 +44714,7 @@ } }, { - "accuracy": 0.9938594889827073, + "accuracy": 0.9938576300628483, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -44757,7 +44757,7 @@ } }, { - "accuracy": 0.9942582775838673, + "accuracy": 0.994258991908282, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -44806,7 +44806,7 @@ } }, { - "accuracy": 0.9951435467228293, + "accuracy": 0.9951435327529907, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -44852,7 +44852,7 @@ } }, { - "accuracy": 0.9983168488834053, + "accuracy": 0.9983171256026253, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -44894,7 +44894,7 @@ ], "model.layers.23.self_attn": [ { - "accuracy": 0.9002235606312752, + "accuracy": 0.9000980034470558, "total_bits": 30308928, "q_proj": { "group_size": { @@ -44958,7 +44958,7 @@ } }, { - "accuracy": 0.9065486118197441, + "accuracy": 0.9066225737333298, "total_bits": 31455808, "q_proj": { "group_size": { @@ -45022,7 +45022,7 @@ } }, { - "accuracy": 0.918041318655014, + "accuracy": 0.9180461913347244, "total_bits": 33412832, "q_proj": { "group_size": { @@ -45086,7 +45086,7 @@ } }, { - "accuracy": 0.9384324066340923, + "accuracy": 0.9384226053953171, "total_bits": 37983200, "q_proj": { "group_size": { @@ -45150,7 +45150,7 @@ } }, { - "accuracy": 0.9503814466297626, + "accuracy": 0.950372826308012, "total_bits": 44838176, "q_proj": { "group_size": { @@ -45214,7 +45214,7 @@ } }, { - "accuracy": 0.9512348845601082, + "accuracy": 0.9512467607855797, "total_bits": 44912768, "q_proj": { "group_size": { @@ -45278,7 +45278,7 @@ } }, { - "accuracy": 0.970819503068924, + "accuracy": 0.9708252772688866, "total_bits": 57355552, "q_proj": { "group_size": { @@ -45330,7 +45330,7 @@ } }, { - "accuracy": 0.9717874713242054, + "accuracy": 0.9717806540429592, "total_bits": 57430144, "q_proj": { "group_size": { @@ -45382,7 +45382,7 @@ } }, { - "accuracy": 0.9738185089081526, + "accuracy": 0.9737685434520245, "total_bits": 57950464, "q_proj": { "group_size": { @@ -45434,7 +45434,7 @@ } }, { - "accuracy": 0.9751602187752724, + "accuracy": 0.9751146398484707, "total_bits": 58692736, "q_proj": { "group_size": { @@ -45486,7 +45486,7 @@ } }, { - "accuracy": 0.9751874227076769, + "accuracy": 0.9751594942063093, "total_bits": 59068544, "q_proj": { "group_size": { @@ -45550,7 +45550,7 @@ } }, { - "accuracy": 0.9768790081143379, + "accuracy": 0.9768443964421749, "total_bits": 59588864, "q_proj": { "group_size": { @@ -45614,7 +45614,7 @@ } }, { - "accuracy": 0.9792286064475775, + "accuracy": 0.9791899565607309, "total_bits": 61536832, "q_proj": { "group_size": { @@ -45675,7 +45675,7 @@ } }, { - "accuracy": 0.9807712472975254, + "accuracy": 0.9807248823344707, "total_bits": 62354560, "q_proj": { "group_size": { @@ -45736,7 +45736,7 @@ } }, { - "accuracy": 0.9885837156325579, + "accuracy": 0.98859893810004, "total_bits": 75246880, "q_proj": { "group_size": { @@ -45797,7 +45797,7 @@ } }, { - "accuracy": 0.9903419725596905, + "accuracy": 0.9903517542406917, "total_bits": 76510336, "q_proj": { "group_size": { @@ -45858,7 +45858,7 @@ } }, { - "accuracy": 0.9923813547939062, + "accuracy": 0.9923766339197755, "total_bits": 85667104, "q_proj": { "group_size": { @@ -45910,7 +45910,7 @@ } }, { - "accuracy": 0.9948961278423667, + "accuracy": 0.9948966847732663, "total_bits": 91722880, "q_proj": { "group_size": { @@ -45962,7 +45962,7 @@ } }, { - "accuracy": 0.9980002944357693, + "accuracy": 0.9980048832949251, "total_bits": 113978656, "q_proj": { "group_size": { @@ -46016,7 +46016,7 @@ ], "model.layers.23.mlp": [ { - "accuracy": 0.9183400496840477, + "accuracy": 0.9182481989264488, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -46068,7 +46068,7 @@ } }, { - "accuracy": 0.9207602441310883, + "accuracy": 0.9207983389496803, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -46120,7 +46120,7 @@ } }, { - "accuracy": 0.9304101318120956, + "accuracy": 0.9304468557238579, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -46169,7 +46169,7 @@ } }, { - "accuracy": 0.9335396587848663, + "accuracy": 0.9335753992199898, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -46218,7 +46218,7 @@ } }, { - "accuracy": 0.95809031650424, + "accuracy": 0.9580657295882702, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -46270,7 +46270,7 @@ } }, { - "accuracy": 0.9615879841148853, + "accuracy": 0.9615712836384773, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -46322,7 +46322,7 @@ } }, { - "accuracy": 0.9659944400191307, + "accuracy": 0.9659555852413177, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -46371,7 +46371,7 @@ } }, { - "accuracy": 0.9778957311064005, + "accuracy": 0.9778776671737432, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -46414,7 +46414,7 @@ } }, { - "accuracy": 0.9797942880541086, + "accuracy": 0.9797797352075577, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -46457,7 +46457,7 @@ } }, { - "accuracy": 0.9784904830157757, + "accuracy": 0.9784805551171303, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -46509,7 +46509,7 @@ } }, { - "accuracy": 0.9812441393733025, + "accuracy": 0.9812456294894218, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -46561,7 +46561,7 @@ } }, { - "accuracy": 0.9889877140522003, + "accuracy": 0.9889850988984108, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -46613,7 +46613,7 @@ } }, { - "accuracy": 0.9906198102980852, + "accuracy": 0.9906222112476826, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -46665,7 +46665,7 @@ } }, { - "accuracy": 0.9938806621357799, + "accuracy": 0.9938766132108867, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -46708,7 +46708,7 @@ } }, { - "accuracy": 0.9943293896503747, + "accuracy": 0.9943245043978095, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -46757,7 +46757,7 @@ } }, { - "accuracy": 0.9952116911299527, + "accuracy": 0.9952051844447851, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -46803,7 +46803,7 @@ } }, { - "accuracy": 0.9983016916085035, + "accuracy": 0.9983023661188781, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -46845,7 +46845,7 @@ ], "model.layers.24.self_attn": [ { - "accuracy": 0.9000411629676819, + "accuracy": 0.9000758603215218, "total_bits": 30308928, "q_proj": { "group_size": { @@ -46909,7 +46909,7 @@ } }, { - "accuracy": 0.9092481806874275, + "accuracy": 0.9089899435639381, "total_bits": 31455808, "q_proj": { "group_size": { @@ -46973,7 +46973,7 @@ } }, { - "accuracy": 0.9228631407022476, + "accuracy": 0.9227519929409027, "total_bits": 33412832, "q_proj": { "group_size": { @@ -47037,7 +47037,7 @@ } }, { - "accuracy": 0.9451075606048107, + "accuracy": 0.9450107850134373, "total_bits": 37983200, "q_proj": { "group_size": { @@ -47101,7 +47101,7 @@ } }, { - "accuracy": 0.9511439241468906, + "accuracy": 0.9511034078896046, "total_bits": 44838176, "q_proj": { "group_size": { @@ -47165,7 +47165,7 @@ } }, { - "accuracy": 0.9521462395787239, + "accuracy": 0.9521391801536083, "total_bits": 44912768, "q_proj": { "group_size": { @@ -47229,7 +47229,7 @@ } }, { - "accuracy": 0.9704149011522532, + "accuracy": 0.9703994449228048, "total_bits": 57355552, "q_proj": { "group_size": { @@ -47281,7 +47281,7 @@ } }, { - "accuracy": 0.9715412594377995, + "accuracy": 0.9715345222502947, "total_bits": 57430144, "q_proj": { "group_size": { @@ -47333,7 +47333,7 @@ } }, { - "accuracy": 0.9736700374633074, + "accuracy": 0.9736871141940355, "total_bits": 57950464, "q_proj": { "group_size": { @@ -47385,7 +47385,7 @@ } }, { - "accuracy": 0.9749877713620663, + "accuracy": 0.9749988541007042, "total_bits": 58692736, "q_proj": { "group_size": { @@ -47437,7 +47437,7 @@ } }, { - "accuracy": 0.9757188335061073, + "accuracy": 0.9757179487496614, "total_bits": 59068544, "q_proj": { "group_size": { @@ -47501,7 +47501,7 @@ } }, { - "accuracy": 0.9774507116526365, + "accuracy": 0.9774703308939934, "total_bits": 59588864, "q_proj": { "group_size": { @@ -47565,7 +47565,7 @@ } }, { - "accuracy": 0.9802383426576853, + "accuracy": 0.9802435729652643, "total_bits": 61536832, "q_proj": { "group_size": { @@ -47626,7 +47626,7 @@ } }, { - "accuracy": 0.9814821984618902, + "accuracy": 0.9814784899353981, "total_bits": 62354560, "q_proj": { "group_size": { @@ -47687,7 +47687,7 @@ } }, { - "accuracy": 0.9890236798673868, + "accuracy": 0.9890194535255432, "total_bits": 75246880, "q_proj": { "group_size": { @@ -47748,7 +47748,7 @@ } }, { - "accuracy": 0.9906842615455389, + "accuracy": 0.9906802931800485, "total_bits": 76510336, "q_proj": { "group_size": { @@ -47809,7 +47809,7 @@ } }, { - "accuracy": 0.9923012899234891, + "accuracy": 0.992301972117275, "total_bits": 85667104, "q_proj": { "group_size": { @@ -47861,7 +47861,7 @@ } }, { - "accuracy": 0.9948786222375929, + "accuracy": 0.9948805985040963, "total_bits": 91722880, "q_proj": { "group_size": { @@ -47913,7 +47913,7 @@ } }, { - "accuracy": 0.997995373327285, + "accuracy": 0.9979949889238924, "total_bits": 113978656, "q_proj": { "group_size": { @@ -47967,7 +47967,7 @@ ], "model.layers.24.mlp": [ { - "accuracy": 0.92559964209795, + "accuracy": 0.9256472215056419, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -48019,7 +48019,7 @@ } }, { - "accuracy": 0.9278130382299423, + "accuracy": 0.9277902320027351, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -48071,7 +48071,7 @@ } }, { - "accuracy": 0.9358445852994919, + "accuracy": 0.9358237609267235, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -48120,7 +48120,7 @@ } }, { - "accuracy": 0.9386767745018005, + "accuracy": 0.9386411532759666, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -48169,7 +48169,7 @@ } }, { - "accuracy": 0.9618233405053616, + "accuracy": 0.9618184603750706, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -48221,7 +48221,7 @@ } }, { - "accuracy": 0.9650960303843021, + "accuracy": 0.9650743305683136, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -48273,7 +48273,7 @@ } }, { - "accuracy": 0.9686915799975395, + "accuracy": 0.9686741307377815, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -48322,7 +48322,7 @@ } }, { - "accuracy": 0.9795693270862103, + "accuracy": 0.9795593917369843, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -48365,7 +48365,7 @@ } }, { - "accuracy": 0.9815245009958744, + "accuracy": 0.9815181270241737, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -48408,7 +48408,7 @@ } }, { - "accuracy": 0.9804772697389126, + "accuracy": 0.9804782513529062, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -48460,7 +48460,7 @@ } }, { - "accuracy": 0.982961056753993, + "accuracy": 0.9829648770391941, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -48512,7 +48512,7 @@ } }, { - "accuracy": 0.9900317844003439, + "accuracy": 0.9900405630469322, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -48564,7 +48564,7 @@ } }, { - "accuracy": 0.9914947487413883, + "accuracy": 0.9914911538362503, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -48616,7 +48616,7 @@ } }, { - "accuracy": 0.994380182120949, + "accuracy": 0.9943781793117523, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -48659,7 +48659,7 @@ } }, { - "accuracy": 0.9948879033327103, + "accuracy": 0.9948839037679136, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -48708,7 +48708,7 @@ } }, { - "accuracy": 0.9956484977155924, + "accuracy": 0.9956428701989353, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -48754,7 +48754,7 @@ } }, { - "accuracy": 0.9983874682802707, + "accuracy": 0.9983870283467695, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -48796,7 +48796,7 @@ ], "model.layers.25.self_attn": [ { - "accuracy": 0.9237724244594574, + "accuracy": 0.923765018582344, "total_bits": 30308928, "q_proj": { "group_size": { @@ -48860,7 +48860,7 @@ } }, { - "accuracy": 0.9273674637079239, + "accuracy": 0.9273531660437584, "total_bits": 31455808, "q_proj": { "group_size": { @@ -48924,7 +48924,7 @@ } }, { - "accuracy": 0.9392730295658112, + "accuracy": 0.9392255060374737, "total_bits": 33412832, "q_proj": { "group_size": { @@ -48988,7 +48988,7 @@ } }, { - "accuracy": 0.9596496224403381, + "accuracy": 0.9596178978681564, "total_bits": 37983200, "q_proj": { "group_size": { @@ -49052,7 +49052,7 @@ } }, { - "accuracy": 0.9613975808024406, + "accuracy": 0.9613914974033833, "total_bits": 44838176, "q_proj": { "group_size": { @@ -49116,7 +49116,7 @@ } }, { - "accuracy": 0.9620046243071556, + "accuracy": 0.962019756436348, "total_bits": 44912768, "q_proj": { "group_size": { @@ -49180,7 +49180,7 @@ } }, { - "accuracy": 0.9788154270499945, + "accuracy": 0.9787868969142437, "total_bits": 57355552, "q_proj": { "group_size": { @@ -49232,7 +49232,7 @@ } }, { - "accuracy": 0.9793564602732658, + "accuracy": 0.9793211314827204, "total_bits": 57430144, "q_proj": { "group_size": { @@ -49284,7 +49284,7 @@ } }, { - "accuracy": 0.980896295979619, + "accuracy": 0.9808873273432255, "total_bits": 57950464, "q_proj": { "group_size": { @@ -49336,7 +49336,7 @@ } }, { - "accuracy": 0.9818288143724203, + "accuracy": 0.9818263165652752, "total_bits": 58692736, "q_proj": { "group_size": { @@ -49388,7 +49388,7 @@ } }, { - "accuracy": 0.9807142727077007, + "accuracy": 0.9806932359933853, "total_bits": 59068544, "q_proj": { "group_size": { @@ -49452,7 +49452,7 @@ } }, { - "accuracy": 0.9820475969463587, + "accuracy": 0.9820628892630339, "total_bits": 59588864, "q_proj": { "group_size": { @@ -49516,7 +49516,7 @@ } }, { - "accuracy": 0.9845238020643592, + "accuracy": 0.9845330221578479, "total_bits": 61536832, "q_proj": { "group_size": { @@ -49577,7 +49577,7 @@ } }, { - "accuracy": 0.9857827695086598, + "accuracy": 0.9857928091660142, "total_bits": 62354560, "q_proj": { "group_size": { @@ -49638,7 +49638,7 @@ } }, { - "accuracy": 0.9914994724094868, + "accuracy": 0.9915020493790507, "total_bits": 75246880, "q_proj": { "group_size": { @@ -49699,7 +49699,7 @@ } }, { - "accuracy": 0.9928717431612313, + "accuracy": 0.9928759732283652, "total_bits": 76510336, "q_proj": { "group_size": { @@ -49760,7 +49760,7 @@ } }, { - "accuracy": 0.9944902677088976, + "accuracy": 0.9944881168194115, "total_bits": 85667104, "q_proj": { "group_size": { @@ -49812,7 +49812,7 @@ } }, { - "accuracy": 0.9964965777471662, + "accuracy": 0.9964962496887892, "total_bits": 91722880, "q_proj": { "group_size": { @@ -49864,7 +49864,7 @@ } }, { - "accuracy": 0.9985775835812092, + "accuracy": 0.998579174396582, "total_bits": 113978656, "q_proj": { "group_size": { @@ -49918,7 +49918,7 @@ ], "model.layers.25.mlp": [ { - "accuracy": 0.9394260719418526, + "accuracy": 0.9394301772117615, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -49970,7 +49970,7 @@ } }, { - "accuracy": 0.9412130787968636, + "accuracy": 0.9411459490656853, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -50022,7 +50022,7 @@ } }, { - "accuracy": 0.9476935565471649, + "accuracy": 0.9477050378918648, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -50071,7 +50071,7 @@ } }, { - "accuracy": 0.9500809907913208, + "accuracy": 0.9500880539417267, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -50120,7 +50120,7 @@ } }, { - "accuracy": 0.9688832182437181, + "accuracy": 0.96889528632164, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -50172,7 +50172,7 @@ } }, { - "accuracy": 0.9714504927396774, + "accuracy": 0.9714055731892586, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -50224,7 +50224,7 @@ } }, { - "accuracy": 0.9744708389043808, + "accuracy": 0.9744217358529568, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -50273,7 +50273,7 @@ } }, { - "accuracy": 0.9832967892289162, + "accuracy": 0.9833166711032391, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -50316,7 +50316,7 @@ } }, { - "accuracy": 0.9849021304398775, + "accuracy": 0.9849346242845058, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -50359,7 +50359,7 @@ } }, { - "accuracy": 0.9840464647859335, + "accuracy": 0.9840444251894951, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -50411,7 +50411,7 @@ } }, { - "accuracy": 0.9860300803557038, + "accuracy": 0.9860321069136262, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -50463,7 +50463,7 @@ } }, { - "accuracy": 0.9918431174010038, + "accuracy": 0.9918465595692396, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -50515,7 +50515,7 @@ } }, { - "accuracy": 0.9930016789585352, + "accuracy": 0.9929989366792142, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -50567,7 +50567,7 @@ } }, { - "accuracy": 0.9954412854276597, + "accuracy": 0.9954434814862907, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -50610,7 +50610,7 @@ } }, { - "accuracy": 0.9958284390158951, + "accuracy": 0.9958340055309236, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -50659,7 +50659,7 @@ } }, { - "accuracy": 0.9965037242509425, + "accuracy": 0.9965090290643275, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -50705,7 +50705,7 @@ } }, { - "accuracy": 0.9987250744597986, + "accuracy": 0.9987249359255657, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -50750,7 +50750,7 @@ }, "strategy": { "model.layers.0.self_attn": { - "accuracy": 0.9889282267540693, + "accuracy": 0.9889282938092947, "total_bits": 59588864, "q_proj": { "group_size": { @@ -50814,81 +50814,78 @@ } }, "model.layers.0.mlp": { - "accuracy": 0.9745050817728043, - "total_bits": 277459200, + "accuracy": 0.971364825963974, + "total_bits": 267653376, "gate_proj": { "group_size": { - "5": 32, "4": 32 }, "bits": [ - 5, 4 ], "bits_prop": [ - 0.1, - 0.9 + 1 ], "scale_bits": 4 }, "up_proj": { "group_size": { - "5": 32, "4": 32 }, "bits": [ - 5, 4 ], "bits_prop": [ - 0.25, - 0.75 + 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "8": 32, - "5": 32, "4": 32 }, "bits": [ 8, - 5, 4 ], "bits_prop": [ 0.05, - 0.1, - 0.85 + 0.95 ], "scale_bits": 4 } }, "model.layers.1.self_attn": { - "accuracy": 0.9933391944505274, - "total_bits": 85667104, + "accuracy": 0.9890200421214104, + "total_bits": 75246880, "q_proj": { "group_size": { - "6": 128 + "6": 128, + "5": 128 }, "bits": [ - 6 + 6, + 5 ], "bits_prop": [ - 1 + 0.1, + 0.9 ], "scale_bits": 4 }, "k_proj": { "group_size": { - "6": 128 + "6": 128, + "5": 128 }, "bits": [ - 6 + 6, + 5 ], "bits_prop": [ - 1 + 0.1, + 0.9 ], "scale_bits": 4 }, @@ -50906,19 +50903,22 @@ }, "o_proj": { "group_size": { - "6": 128 + "6": 128, + "5": 128 }, "bits": [ - 6 + 6, + 5 ], "bits_prop": [ - 1 + 0.1, + 0.9 ], "scale_bits": 4 } }, "model.layers.1.mlp": { - "accuracy": 0.9747709520161152, + "accuracy": 0.9747742302715778, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -50970,12 +50970,12 @@ } }, "model.layers.2.self_attn": { - "accuracy": 0.9874144792556763, - "total_bits": 76510336, + "accuracy": 0.9853210505098104, + "total_bits": 75246880, "q_proj": { "group_size": { - "6": 32, - "5": 32 + "6": 128, + "5": 128 }, "bits": [ 6, @@ -50989,8 +50989,8 @@ }, "k_proj": { "group_size": { - "6": 32, - "5": 32 + "6": 128, + "5": 128 }, "bits": [ 6, @@ -51004,7 +51004,7 @@ }, "v_proj": { "group_size": { - "6": 32 + "6": 128 }, "bits": [ 6 @@ -51016,8 +51016,8 @@ }, "o_proj": { "group_size": { - "6": 32, - "5": 32 + "6": 128, + "5": 128 }, "bits": [ 6, @@ -51031,7 +51031,7 @@ } }, "model.layers.2.mlp": { - "accuracy": 0.966707780957222, + "accuracy": 0.9667122215032578, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -51083,38 +51083,44 @@ } }, "model.layers.3.self_attn": { - "accuracy": 0.994609275367111, - "total_bits": 91722880, + "accuracy": 0.9891604781150818, + "total_bits": 76510336, "q_proj": { "group_size": { - "6": 32 + "6": 32, + "5": 32 }, "bits": [ - 6 + 6, + 5 ], "bits_prop": [ - 1 + 0.1, + 0.9 ], "scale_bits": 4 }, "k_proj": { "group_size": { - "6": 32 + "6": 32, + "5": 32 }, "bits": [ - 6 + 6, + 5 ], "bits_prop": [ - 1 + 0.1, + 0.9 ], "scale_bits": 4 }, "v_proj": { "group_size": { - "8": 32 + "6": 32 }, "bits": [ - 8 + 6 ], "bits_prop": [ 1 @@ -51123,19 +51129,22 @@ }, "o_proj": { "group_size": { - "6": 32 + "6": 32, + "5": 32 }, "bits": [ - 6 + 6, + 5 ], "bits_prop": [ - 1 + 0.1, + 0.9 ], "scale_bits": 4 } }, "model.layers.3.mlp": { - "accuracy": 0.9798263423144817, + "accuracy": 0.979815537109971, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -51187,11 +51196,11 @@ } }, "model.layers.4.self_attn": { - "accuracy": 0.9939752677455544, - "total_bits": 91722880, + "accuracy": 0.9909781841561198, + "total_bits": 85667104, "q_proj": { "group_size": { - "6": 32 + "6": 128 }, "bits": [ 6 @@ -51203,7 +51212,7 @@ }, "k_proj": { "group_size": { - "6": 32 + "6": 128 }, "bits": [ 6 @@ -51215,10 +51224,10 @@ }, "v_proj": { "group_size": { - "8": 32 + "6": 128 }, "bits": [ - 8 + 6 ], "bits_prop": [ 1 @@ -51227,7 +51236,7 @@ }, "o_proj": { "group_size": { - "6": 32 + "6": 128 }, "bits": [ 6 @@ -51239,7 +51248,7 @@ } }, "model.layers.4.mlp": { - "accuracy": 0.9814701918512583, + "accuracy": 0.9814739804714918, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -51291,12 +51300,12 @@ } }, "model.layers.5.self_attn": { - "accuracy": 0.9886289816349745, - "total_bits": 76510336, + "accuracy": 0.9866563268005848, + "total_bits": 75246880, "q_proj": { "group_size": { - "6": 32, - "5": 32 + "6": 128, + "5": 128 }, "bits": [ 6, @@ -51310,8 +51319,8 @@ }, "k_proj": { "group_size": { - "6": 32, - "5": 32 + "6": 128, + "5": 128 }, "bits": [ 6, @@ -51325,7 +51334,7 @@ }, "v_proj": { "group_size": { - "6": 32 + "6": 128 }, "bits": [ 6 @@ -51337,8 +51346,8 @@ }, "o_proj": { "group_size": { - "6": 32, - "5": 32 + "6": 128, + "5": 128 }, "bits": [ 6, @@ -51352,7 +51361,7 @@ } }, "model.layers.5.mlp": { - "accuracy": 0.9695848729461432, + "accuracy": 0.9695756994187832, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -51404,35 +51413,41 @@ } }, "model.layers.6.self_attn": { - "accuracy": 0.9920955216512084, - "total_bits": 85667104, + "accuracy": 0.9893340524286032, + "total_bits": 76510336, "q_proj": { "group_size": { - "6": 128 + "6": 32, + "5": 32 }, "bits": [ - 6 + 6, + 5 ], "bits_prop": [ - 1 + 0.1, + 0.9 ], "scale_bits": 4 }, "k_proj": { "group_size": { - "6": 128 + "6": 32, + "5": 32 }, "bits": [ - 6 + 6, + 5 ], "bits_prop": [ - 1 + 0.1, + 0.9 ], "scale_bits": 4 }, "v_proj": { "group_size": { - "6": 128 + "6": 32 }, "bits": [ 6 @@ -51444,19 +51459,22 @@ }, "o_proj": { "group_size": { - "6": 128 + "6": 32, + "5": 32 }, "bits": [ - 6 + 6, + 5 ], "bits_prop": [ - 1 + 0.1, + 0.9 ], "scale_bits": 4 } }, "model.layers.6.mlp": { - "accuracy": 0.9682779163122177, + "accuracy": 0.9682538993656635, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -51508,41 +51526,35 @@ } }, "model.layers.7.self_attn": { - "accuracy": 0.9881619503721595, - "total_bits": 76510336, + "accuracy": 0.9907346172258258, + "total_bits": 85667104, "q_proj": { "group_size": { - "6": 32, - "5": 32 + "6": 128 }, "bits": [ - 6, - 5 + 6 ], "bits_prop": [ - 0.1, - 0.9 + 1 ], "scale_bits": 4 }, "k_proj": { "group_size": { - "6": 32, - "5": 32 + "6": 128 }, "bits": [ - 6, - 5 + 6 ], "bits_prop": [ - 0.1, - 0.9 + 1 ], "scale_bits": 4 }, "v_proj": { "group_size": { - "6": 32 + "6": 128 }, "bits": [ 6 @@ -51554,22 +51566,19 @@ }, "o_proj": { "group_size": { - "6": 32, - "5": 32 + "6": 128 }, "bits": [ - 6, - 5 + 6 ], "bits_prop": [ - 0.1, - 0.9 + 1 ], "scale_bits": 4 } }, "model.layers.7.mlp": { - "accuracy": 0.9720455892384052, + "accuracy": 0.9720457717776299, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -51612,38 +51621,44 @@ } }, "model.layers.8.self_attn": { - "accuracy": 0.9819746408611536, - "total_bits": 58692736, + "accuracy": 0.9848749991506338, + "total_bits": 62354560, "q_proj": { "group_size": { + "5": 32, "4": 32 }, "bits": [ + 5, 4 ], "bits_prop": [ - 1 + 0.1, + 0.9 ], "scale_bits": 4 }, "k_proj": { "group_size": { + "5": 32, "4": 32 }, "bits": [ + 5, 4 ], "bits_prop": [ - 1 + 0.1, + 0.9 ], "scale_bits": 4 }, "v_proj": { "group_size": { - "4": 32 + "5": 32 }, "bits": [ - 4 + 5 ], "bits_prop": [ 1 @@ -51652,71 +51667,65 @@ }, "o_proj": { "group_size": { + "5": 32, "4": 32 }, "bits": [ + 5, 4 ], "bits_prop": [ - 1 + 0.1, + 0.9 ], "scale_bits": 4 } }, "model.layers.8.mlp": { - "accuracy": 0.972480921074748, - "total_bits": 277459200, + "accuracy": 0.9705599583685398, + "total_bits": 267653376, "gate_proj": { "group_size": { - "5": 32, "4": 32 }, "bits": [ - 5, 4 ], "bits_prop": [ - 0.1, - 0.9 + 1 ], "scale_bits": 4 }, "up_proj": { "group_size": { - "5": 32, "4": 32 }, "bits": [ - 5, 4 ], "bits_prop": [ - 0.25, - 0.75 + 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "8": 32, - "5": 32, "4": 32 }, "bits": [ 8, - 5, 4 ], "bits_prop": [ 0.05, - 0.1, - 0.85 + 0.95 ], "scale_bits": 4 } }, "model.layers.9.self_attn": { - "accuracy": 0.9902575109153986, + "accuracy": 0.9902483588084579, "total_bits": 75246880, "q_proj": { "group_size": { @@ -51777,90 +51786,87 @@ } }, "model.layers.9.mlp": { - "accuracy": 0.9767951015383005, - "total_bits": 277459200, + "accuracy": 0.975043885409832, + "total_bits": 267653376, "gate_proj": { "group_size": { - "5": 32, "4": 32 }, "bits": [ - 5, 4 ], "bits_prop": [ - 0.1, - 0.9 + 1 ], "scale_bits": 4 }, "up_proj": { "group_size": { - "5": 32, "4": 32 }, "bits": [ - 5, 4 ], "bits_prop": [ - 0.25, - 0.75 + 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "8": 32, - "5": 32, "4": 32 }, "bits": [ 8, - 5, 4 ], "bits_prop": [ 0.05, - 0.1, - 0.85 + 0.95 ], "scale_bits": 4 } }, "model.layers.10.self_attn": { - "accuracy": 0.982187744230032, - "total_bits": 58692736, + "accuracy": 0.9842821806669235, + "total_bits": 61536832, "q_proj": { "group_size": { - "4": 32 + "5": 64, + "4": 64 }, "bits": [ + 5, 4 ], "bits_prop": [ - 1 + 0.1, + 0.9 ], "scale_bits": 4 }, "k_proj": { "group_size": { - "4": 32 + "5": 64, + "4": 64 }, "bits": [ + 5, 4 ], "bits_prop": [ - 1 + 0.1, + 0.9 ], "scale_bits": 4 }, "v_proj": { "group_size": { - "4": 32 + "5": 64 }, "bits": [ - 4 + 5 ], "bits_prop": [ 1 @@ -51869,23 +51875,26 @@ }, "o_proj": { "group_size": { - "4": 32 + "5": 64, + "4": 64 }, "bits": [ + 5, 4 ], "bits_prop": [ - 1 + 0.1, + 0.9 ], "scale_bits": 4 } }, "model.layers.10.mlp": { - "accuracy": 0.975079670548439, - "total_bits": 263770272, + "accuracy": 0.9773098323494196, + "total_bits": 267653376, "gate_proj": { "group_size": { - "4": 128 + "4": 32 }, "bits": [ 4 @@ -51910,7 +51919,7 @@ "down_proj": { "group_size": { "8": 32, - "4": 128 + "4": 32 }, "bits": [ 8, @@ -51924,44 +51933,38 @@ } }, "model.layers.11.self_attn": { - "accuracy": 0.9868558822199702, - "total_bits": 62354560, + "accuracy": 0.9834803491830826, + "total_bits": 57950464, "q_proj": { "group_size": { - "5": 32, - "4": 32 + "4": 64 }, "bits": [ - 5, 4 ], "bits_prop": [ - 0.1, - 0.9 + 1 ], "scale_bits": 4 }, "k_proj": { "group_size": { - "5": 32, - "4": 32 + "4": 64 }, "bits": [ - 5, 4 ], "bits_prop": [ - 0.1, - 0.9 + 1 ], "scale_bits": 4 }, "v_proj": { "group_size": { - "5": 32 + "4": 32 }, "bits": [ - 5 + 4 ], "bits_prop": [ 1 @@ -51970,26 +51973,23 @@ }, "o_proj": { "group_size": { - "5": 32, - "4": 32 + "4": 64 }, "bits": [ - 5, 4 ], "bits_prop": [ - 0.1, - 0.9 + 1 ], "scale_bits": 4 } }, "model.layers.11.mlp": { - "accuracy": 0.9810750614851713, - "total_bits": 267653376, + "accuracy": 0.9791763704270124, + "total_bits": 263770272, "gate_proj": { "group_size": { - "4": 32 + "4": 128 }, "bits": [ 4 @@ -52014,7 +52014,7 @@ "down_proj": { "group_size": { "8": 32, - "4": 32 + "4": 128 }, "bits": [ 8, @@ -52028,41 +52028,35 @@ } }, "model.layers.12.self_attn": { - "accuracy": 0.9914399096742272, - "total_bits": 76510336, + "accuracy": 0.9938817266374826, + "total_bits": 85667104, "q_proj": { "group_size": { - "6": 32, - "5": 32 + "6": 128 }, "bits": [ - 6, - 5 + 6 ], "bits_prop": [ - 0.1, - 0.9 + 1 ], "scale_bits": 4 }, "k_proj": { "group_size": { - "6": 32, - "5": 32 + "6": 128 }, "bits": [ - 6, - 5 + 6 ], "bits_prop": [ - 0.1, - 0.9 + 1 ], "scale_bits": 4 }, "v_proj": { "group_size": { - "6": 32 + "6": 128 }, "bits": [ 6 @@ -52074,22 +52068,19 @@ }, "o_proj": { "group_size": { - "6": 32, - "5": 32 + "6": 128 }, "bits": [ - 6, - 5 + 6 ], "bits_prop": [ - 0.1, - 0.9 + 1 ], "scale_bits": 4 } }, "model.layers.12.mlp": { - "accuracy": 0.9771162606775761, + "accuracy": 0.9771051350980997, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -52132,16 +52123,16 @@ } }, "model.layers.13.self_attn": { - "accuracy": 0.9835209306329489, - "total_bits": 62354560, + "accuracy": 0.9917331263422966, + "total_bits": 76510336, "q_proj": { "group_size": { - "5": 32, - "4": 32 + "6": 32, + "5": 32 }, "bits": [ - 5, - 4 + 6, + 5 ], "bits_prop": [ 0.1, @@ -52151,12 +52142,12 @@ }, "k_proj": { "group_size": { - "5": 32, - "4": 32 + "6": 32, + "5": 32 }, "bits": [ - 5, - 4 + 6, + 5 ], "bits_prop": [ 0.1, @@ -52166,10 +52157,10 @@ }, "v_proj": { "group_size": { - "5": 32 + "6": 32 }, "bits": [ - 5 + 6 ], "bits_prop": [ 1 @@ -52178,12 +52169,12 @@ }, "o_proj": { "group_size": { - "5": 32, - "4": 32 + "6": 32, + "5": 32 }, "bits": [ - 5, - 4 + 6, + 5 ], "bits_prop": [ 0.1, @@ -52193,7 +52184,7 @@ } }, "model.layers.13.mlp": { - "accuracy": 0.976823752745986, + "accuracy": 0.9768188558518887, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -52236,35 +52227,41 @@ } }, "model.layers.14.self_attn": { - "accuracy": 0.992810903582722, - "total_bits": 85667104, + "accuracy": 0.9902975112199783, + "total_bits": 76510336, "q_proj": { "group_size": { - "6": 128 + "6": 32, + "5": 32 }, "bits": [ - 6 + 6, + 5 ], "bits_prop": [ - 1 + 0.1, + 0.9 ], "scale_bits": 4 }, "k_proj": { "group_size": { - "6": 128 + "6": 32, + "5": 32 }, "bits": [ - 6 + 6, + 5 ], "bits_prop": [ - 1 + 0.1, + 0.9 ], "scale_bits": 4 }, "v_proj": { "group_size": { - "6": 128 + "6": 32 }, "bits": [ 6 @@ -52276,23 +52273,26 @@ }, "o_proj": { "group_size": { - "6": 128 + "6": 32, + "5": 32 }, "bits": [ - 6 + 6, + 5 ], "bits_prop": [ - 1 + 0.1, + 0.9 ], "scale_bits": 4 } }, "model.layers.14.mlp": { - "accuracy": 0.9745946303009987, - "total_bits": 263770272, + "accuracy": 0.9768539238721132, + "total_bits": 267653376, "gate_proj": { "group_size": { - "4": 128 + "4": 32 }, "bits": [ 4 @@ -52317,7 +52317,7 @@ "down_proj": { "group_size": { "8": 32, - "4": 128 + "4": 32 }, "bits": [ 8, @@ -52331,7 +52331,7 @@ } }, "model.layers.15.self_attn": { - "accuracy": 0.987136579118669, + "accuracy": 0.9871381670236588, "total_bits": 76510336, "q_proj": { "group_size": { @@ -52392,59 +52392,50 @@ } }, "model.layers.15.mlp": { - "accuracy": 0.9747244548052549, - "total_bits": 277459200, + "accuracy": 0.9701192807406187, + "total_bits": 263770272, "gate_proj": { "group_size": { - "5": 32, - "4": 32 + "4": 128 }, "bits": [ - 5, 4 ], "bits_prop": [ - 0.1, - 0.9 + 1 ], "scale_bits": 4 }, "up_proj": { "group_size": { - "5": 32, "4": 32 }, "bits": [ - 5, 4 ], "bits_prop": [ - 0.25, - 0.75 + 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "8": 32, - "5": 32, - "4": 32 + "4": 128 }, "bits": [ 8, - 5, 4 ], "bits_prop": [ 0.05, - 0.1, - 0.85 + 0.95 ], "scale_bits": 4 } }, "model.layers.16.self_attn": { - "accuracy": 0.9901310745626688, + "accuracy": 0.9901269851252437, "total_bits": 76510336, "q_proj": { "group_size": { @@ -52505,78 +52496,93 @@ } }, "model.layers.16.mlp": { - "accuracy": 0.97197007201612, - "total_bits": 263770272, + "accuracy": 0.9763544015586376, + "total_bits": 277459200, "gate_proj": { "group_size": { - "4": 128 + "5": 32, + "4": 32 }, "bits": [ + 5, 4 ], "bits_prop": [ - 1 + 0.1, + 0.9 ], "scale_bits": 4 }, "up_proj": { "group_size": { + "5": 32, "4": 32 }, "bits": [ + 5, 4 ], "bits_prop": [ - 1 + 0.25, + 0.75 ], "scale_bits": 4 }, "down_proj": { "group_size": { "8": 32, - "4": 128 + "5": 32, + "4": 32 }, "bits": [ 8, + 5, 4 ], "bits_prop": [ 0.05, - 0.95 + 0.1, + 0.85 ], "scale_bits": 4 } }, "model.layers.17.self_attn": { - "accuracy": 0.9923741007223725, - "total_bits": 85667104, + "accuracy": 0.9897674126550555, + "total_bits": 76510336, "q_proj": { "group_size": { - "6": 128 + "6": 32, + "5": 32 }, "bits": [ - 6 + 6, + 5 ], "bits_prop": [ - 1 + 0.1, + 0.9 ], "scale_bits": 4 }, "k_proj": { "group_size": { - "6": 128 + "6": 32, + "5": 32 }, "bits": [ - 6 + 6, + 5 ], "bits_prop": [ - 1 + 0.1, + 0.9 ], "scale_bits": 4 }, "v_proj": { "group_size": { - "6": 128 + "6": 32 }, "bits": [ 6 @@ -52588,71 +52594,65 @@ }, "o_proj": { "group_size": { - "6": 128 + "6": 32, + "5": 32 }, "bits": [ - 6 + 6, + 5 ], "bits_prop": [ - 1 + 0.1, + 0.9 ], "scale_bits": 4 } }, "model.layers.17.mlp": { - "accuracy": 0.973537739366293, - "total_bits": 277459200, + "accuracy": 0.9712157864123583, + "total_bits": 267653376, "gate_proj": { "group_size": { - "5": 32, "4": 32 }, "bits": [ - 5, 4 ], "bits_prop": [ - 0.1, - 0.9 + 1 ], "scale_bits": 4 }, "up_proj": { "group_size": { - "5": 32, "4": 32 }, "bits": [ - 5, 4 ], "bits_prop": [ - 0.25, - 0.75 + 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "8": 32, - "5": 32, "4": 32 }, "bits": [ 8, - 5, 4 ], "bits_prop": [ 0.05, - 0.1, - 0.85 + 0.95 ], "scale_bits": 4 } }, "model.layers.18.self_attn": { - "accuracy": 0.9899277659133077, + "accuracy": 0.9899393441155553, "total_bits": 76510336, "q_proj": { "group_size": { @@ -52713,68 +52713,59 @@ } }, "model.layers.18.mlp": { - "accuracy": 0.9754060637205839, - "total_bits": 277459200, + "accuracy": 0.9732760544866323, + "total_bits": 267653376, "gate_proj": { "group_size": { - "5": 32, "4": 32 }, "bits": [ - 5, 4 ], "bits_prop": [ - 0.1, - 0.9 + 1 ], "scale_bits": 4 }, "up_proj": { "group_size": { - "5": 32, "4": 32 }, "bits": [ - 5, 4 ], "bits_prop": [ - 0.25, - 0.75 + 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "8": 32, - "5": 32, "4": 32 }, "bits": [ 8, - 5, 4 ], "bits_prop": [ 0.05, - 0.1, - 0.85 + 0.95 ], "scale_bits": 4 } }, "model.layers.19.self_attn": { - "accuracy": 0.9908470399677753, - "total_bits": 76510336, + "accuracy": 0.9817991387099028, + "total_bits": 62354560, "q_proj": { "group_size": { - "6": 32, - "5": 32 + "5": 32, + "4": 32 }, "bits": [ - 6, - 5 + 5, + 4 ], "bits_prop": [ 0.1, @@ -52784,12 +52775,12 @@ }, "k_proj": { "group_size": { - "6": 32, - "5": 32 + "5": 32, + "4": 32 }, "bits": [ - 6, - 5 + 5, + 4 ], "bits_prop": [ 0.1, @@ -52799,10 +52790,10 @@ }, "v_proj": { "group_size": { - "6": 32 + "5": 32 }, "bits": [ - 6 + 5 ], "bits_prop": [ 1 @@ -52811,12 +52802,12 @@ }, "o_proj": { "group_size": { - "6": 32, - "5": 32 + "5": 32, + "4": 32 }, "bits": [ - 6, - 5 + 5, + 4 ], "bits_prop": [ 0.1, @@ -52826,7 +52817,7 @@ } }, "model.layers.19.mlp": { - "accuracy": 0.9751972481608391, + "accuracy": 0.9751956835389137, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -52869,44 +52860,38 @@ } }, "model.layers.20.self_attn": { - "accuracy": 0.9851696165278554, - "total_bits": 62354560, + "accuracy": 0.996142350602895, + "total_bits": 91722880, "q_proj": { "group_size": { - "5": 32, - "4": 32 + "6": 32 }, "bits": [ - 5, - 4 + 6 ], "bits_prop": [ - 0.1, - 0.9 + 1 ], "scale_bits": 4 }, "k_proj": { "group_size": { - "5": 32, - "4": 32 + "6": 32 }, "bits": [ - 5, - 4 + 6 ], "bits_prop": [ - 0.1, - 0.9 + 1 ], "scale_bits": 4 }, "v_proj": { "group_size": { - "5": 32 + "8": 32 }, "bits": [ - 5 + 8 ], "bits_prop": [ 1 @@ -52915,111 +52900,93 @@ }, "o_proj": { "group_size": { - "5": 32, - "4": 32 + "6": 32 }, "bits": [ - 5, - 4 + 6 ], "bits_prop": [ - 0.1, - 0.9 + 1 ], "scale_bits": 4 } }, "model.layers.20.mlp": { - "accuracy": 0.9790751449763775, - "total_bits": 277459200, + "accuracy": 0.9774747285991907, + "total_bits": 267653376, "gate_proj": { "group_size": { - "5": 32, "4": 32 }, "bits": [ - 5, 4 ], "bits_prop": [ - 0.1, - 0.9 + 1 ], "scale_bits": 4 }, "up_proj": { "group_size": { - "5": 32, "4": 32 }, "bits": [ - 5, 4 ], "bits_prop": [ - 0.25, - 0.75 + 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "8": 32, - "5": 32, "4": 32 }, "bits": [ 8, - 5, 4 ], "bits_prop": [ 0.05, - 0.1, - 0.85 + 0.95 ], "scale_bits": 4 } }, "model.layers.21.self_attn": { - "accuracy": 0.9903135802596807, - "total_bits": 75246880, + "accuracy": 0.995899414177984, + "total_bits": 91722880, "q_proj": { "group_size": { - "6": 128, - "5": 128 + "6": 32 }, "bits": [ - 6, - 5 + 6 ], "bits_prop": [ - 0.1, - 0.9 + 1 ], "scale_bits": 4 }, "k_proj": { "group_size": { - "6": 128, - "5": 128 + "6": 32 }, "bits": [ - 6, - 5 + 6 ], "bits_prop": [ - 0.1, - 0.9 + 1 ], "scale_bits": 4 }, "v_proj": { "group_size": { - "6": 128 + "8": 32 }, "bits": [ - 6 + 8 ], "bits_prop": [ 1 @@ -53028,22 +52995,19 @@ }, "o_proj": { "group_size": { - "6": 128, - "5": 128 + "6": 32 }, "bits": [ - 6, - 5 + 6 ], "bits_prop": [ - 0.1, - 0.9 + 1 ], "scale_bits": 4 } }, "model.layers.21.mlp": { - "accuracy": 0.9787649121135473, + "accuracy": 0.9787469636648893, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -53086,11 +53050,11 @@ } }, "model.layers.22.self_attn": { - "accuracy": 0.9924624958075583, - "total_bits": 91722880, + "accuracy": 0.9888947391882539, + "total_bits": 85667104, "q_proj": { "group_size": { - "6": 32 + "6": 128 }, "bits": [ 6 @@ -53102,7 +53066,7 @@ }, "k_proj": { "group_size": { - "6": 32 + "6": 128 }, "bits": [ 6 @@ -53114,10 +53078,10 @@ }, "v_proj": { "group_size": { - "8": 32 + "6": 128 }, "bits": [ - 8 + 6 ], "bits_prop": [ 1 @@ -53126,7 +53090,7 @@ }, "o_proj": { "group_size": { - "6": 32 + "6": 128 }, "bits": [ 6 @@ -53138,50 +53102,59 @@ } }, "model.layers.22.mlp": { - "accuracy": 0.9793780688196421, - "total_bits": 267653376, + "accuracy": 0.9807199090719223, + "total_bits": 277459200, "gate_proj": { "group_size": { + "5": 32, "4": 32 }, "bits": [ + 5, 4 ], "bits_prop": [ - 1 + 0.1, + 0.9 ], "scale_bits": 4 }, "up_proj": { "group_size": { + "5": 32, "4": 32 }, "bits": [ + 5, 4 ], "bits_prop": [ - 1 + 0.25, + 0.75 ], "scale_bits": 4 }, "down_proj": { "group_size": { "8": 32, + "5": 32, "4": 32 }, "bits": [ 8, + 5, 4 ], "bits_prop": [ 0.05, - 0.95 + 0.1, + 0.85 ], "scale_bits": 4 } }, "model.layers.23.self_attn": { - "accuracy": 0.9903419725596905, + "accuracy": 0.9903517542406917, "total_bits": 76510336, "q_proj": { "group_size": { @@ -53242,7 +53215,7 @@ } }, "model.layers.23.mlp": { - "accuracy": 0.9778957311064005, + "accuracy": 0.9778776671737432, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -53285,7 +53258,7 @@ } }, "model.layers.24.self_attn": { - "accuracy": 0.9906842615455389, + "accuracy": 0.9906802931800485, "total_bits": 76510336, "q_proj": { "group_size": { @@ -53346,64 +53319,55 @@ } }, "model.layers.24.mlp": { - "accuracy": 0.982961056753993, - "total_bits": 277459200, + "accuracy": 0.9815181270241737, + "total_bits": 267653376, "gate_proj": { "group_size": { - "5": 32, "4": 32 }, "bits": [ - 5, 4 ], "bits_prop": [ - 0.1, - 0.9 + 1 ], "scale_bits": 4 }, "up_proj": { "group_size": { - "5": 32, "4": 32 }, "bits": [ - 5, 4 ], "bits_prop": [ - 0.25, - 0.75 + 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "8": 32, - "5": 32, "4": 32 }, "bits": [ 8, - 5, 4 ], "bits_prop": [ 0.05, - 0.1, - 0.85 + 0.95 ], "scale_bits": 4 } }, "model.layers.25.self_attn": { - "accuracy": 0.9857827695086598, - "total_bits": 62354560, + "accuracy": 0.9845330221578479, + "total_bits": 61536832, "q_proj": { "group_size": { - "5": 32, - "4": 32 + "5": 64, + "4": 64 }, "bits": [ 5, @@ -53417,8 +53381,8 @@ }, "k_proj": { "group_size": { - "5": 32, - "4": 32 + "5": 64, + "4": 64 }, "bits": [ 5, @@ -53432,7 +53396,7 @@ }, "v_proj": { "group_size": { - "5": 32 + "5": 64 }, "bits": [ 5 @@ -53444,8 +53408,8 @@ }, "o_proj": { "group_size": { - "5": 32, - "4": 32 + "5": 64, + "4": 64 }, "bits": [ 5, @@ -53459,56 +53423,48 @@ } }, "model.layers.25.mlp": { - "accuracy": 0.9688832182437181, - "total_bits": 209798784, + "accuracy": 0.9849346242845058, + "total_bits": 267653376, "gate_proj": { "group_size": { - "4": 128, - "3": 128 + "4": 32 }, "bits": [ - 4, - 3 + 4 ], "bits_prop": [ - 0.1, - 0.9 + 1 ], "scale_bits": 4 }, "up_proj": { "group_size": { - "4": 128, - "3": 128 + "4": 32 }, "bits": [ - 4, - 3 + 4 ], "bits_prop": [ - 0.25, - 0.75 + 1 ], "scale_bits": 4 }, "down_proj": { "group_size": { "8": 32, - "4": 128, - "3": 128 + "4": 32 }, "bits": [ 8, - 4, - 3 + 4 ], "bits_prop": [ 0.05, - 0.1, - 0.85 + 0.95 ], "scale_bits": 4 } } - } + }, + "q_last_module_idx": 54 } \ No newline at end of file diff --git a/measurement.json b/measurement.json index 3a5fa129bd53dc81d75a6da6a9726c7818fef3dd..8be3f6938c3afec312867b9e9cf00f9857f2bad9 100644 --- a/measurement.json +++ b/measurement.json @@ -2,7 +2,7 @@ "measurement": { "model.layers.0.self_attn": [ { - "accuracy": 0.9515928328037262, + "accuracy": 0.9515531063079834, "total_bits": 30308928, "q_proj": { "group_size": { @@ -66,7 +66,7 @@ } }, { - "accuracy": 0.953981950879097, + "accuracy": 0.9539748653769493, "total_bits": 31455808, "q_proj": { "group_size": { @@ -130,7 +130,7 @@ } }, { - "accuracy": 0.9591670855879784, + "accuracy": 0.9591861926019192, "total_bits": 33412832, "q_proj": { "group_size": { @@ -194,7 +194,7 @@ } }, { - "accuracy": 0.9699935875833035, + "accuracy": 0.9699958208948374, "total_bits": 37983200, "q_proj": { "group_size": { @@ -258,7 +258,7 @@ } }, { - "accuracy": 0.976397393271327, + "accuracy": 0.9763969834893942, "total_bits": 44838176, "q_proj": { "group_size": { @@ -322,7 +322,7 @@ } }, { - "accuracy": 0.9767184294760227, + "accuracy": 0.9767143614590168, "total_bits": 44912768, "q_proj": { "group_size": { @@ -386,7 +386,7 @@ } }, { - "accuracy": 0.986974217928946, + "accuracy": 0.9869659701362252, "total_bits": 57355552, "q_proj": { "group_size": { @@ -438,7 +438,7 @@ } }, { - "accuracy": 0.9872219935059547, + "accuracy": 0.9872188298031688, "total_bits": 57430144, "q_proj": { "group_size": { @@ -490,7 +490,7 @@ } }, { - "accuracy": 0.9881364433094859, + "accuracy": 0.9881322896108031, "total_bits": 57950464, "q_proj": { "group_size": { @@ -542,7 +542,7 @@ } }, { - "accuracy": 0.9888399662449956, + "accuracy": 0.9888321068137884, "total_bits": 58692736, "q_proj": { "group_size": { @@ -594,7 +594,7 @@ } }, { - "accuracy": 0.9880855781957507, + "accuracy": 0.9880868950858712, "total_bits": 59068544, "q_proj": { "group_size": { @@ -658,7 +658,7 @@ } }, { - "accuracy": 0.9889282267540693, + "accuracy": 0.9889282938092947, "total_bits": 59588864, "q_proj": { "group_size": { @@ -722,7 +722,7 @@ } }, { - "accuracy": 0.9899680046364665, + "accuracy": 0.9899678928777575, "total_bits": 61536832, "q_proj": { "group_size": { @@ -783,7 +783,7 @@ } }, { - "accuracy": 0.9907521335408092, + "accuracy": 0.9907488031312823, "total_bits": 62354560, "q_proj": { "group_size": { @@ -844,7 +844,7 @@ } }, { - "accuracy": 0.9944849004969001, + "accuracy": 0.9944888786412776, "total_bits": 75246880, "q_proj": { "group_size": { @@ -905,7 +905,7 @@ } }, { - "accuracy": 0.9953417172655463, + "accuracy": 0.9953421992249787, "total_bits": 76510336, "q_proj": { "group_size": { @@ -966,7 +966,7 @@ } }, { - "accuracy": 0.9965903947595507, + "accuracy": 0.9965901931282133, "total_bits": 85667104, "q_proj": { "group_size": { @@ -1018,7 +1018,7 @@ } }, { - "accuracy": 0.9975632540881634, + "accuracy": 0.9975674538873136, "total_bits": 91722880, "q_proj": { "group_size": { @@ -1070,7 +1070,7 @@ } }, { - "accuracy": 0.9990884908474982, + "accuracy": 0.9990892710047774, "total_bits": 113978656, "q_proj": { "group_size": { @@ -1124,7 +1124,7 @@ ], "model.layers.0.mlp": [ { - "accuracy": 0.8881748169660568, + "accuracy": 0.8882661387324333, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -1176,7 +1176,7 @@ } }, { - "accuracy": 0.8919973969459534, + "accuracy": 0.8919670283794403, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -1228,7 +1228,7 @@ } }, { - "accuracy": 0.9069116935133934, + "accuracy": 0.9069375917315483, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -1277,7 +1277,7 @@ } }, { - "accuracy": 0.9124017059803009, + "accuracy": 0.9124317467212677, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -1326,7 +1326,7 @@ } }, { - "accuracy": 0.9430184848606586, + "accuracy": 0.9430231153964996, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -1378,7 +1378,7 @@ } }, { - "accuracy": 0.9480531290173531, + "accuracy": 0.9480909705162048, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -1430,7 +1430,7 @@ } }, { - "accuracy": 0.9544955678284168, + "accuracy": 0.9545128084719181, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -1479,7 +1479,7 @@ } }, { - "accuracy": 0.968281302601099, + "accuracy": 0.9682874381542206, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -1522,7 +1522,7 @@ } }, { - "accuracy": 0.9713601395487785, + "accuracy": 0.971364825963974, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -1565,7 +1565,7 @@ } }, { - "accuracy": 0.9707059450447559, + "accuracy": 0.9707299135625362, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -1617,7 +1617,7 @@ } }, { - "accuracy": 0.9745050817728043, + "accuracy": 0.974507249891758, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -1669,7 +1669,7 @@ } }, { - "accuracy": 0.9848953913897276, + "accuracy": 0.9848882537335157, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -1721,7 +1721,7 @@ } }, { - "accuracy": 0.9872103426605463, + "accuracy": 0.9872004305943847, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -1773,7 +1773,7 @@ } }, { - "accuracy": 0.9911489551886916, + "accuracy": 0.9911568406969309, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -1816,7 +1816,7 @@ } }, { - "accuracy": 0.9923222842626274, + "accuracy": 0.9923209701664746, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -1865,7 +1865,7 @@ } }, { - "accuracy": 0.9939925689250231, + "accuracy": 0.9939967971295118, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -1911,7 +1911,7 @@ } }, { - "accuracy": 0.9975718185305595, + "accuracy": 0.9975708881393075, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -1953,7 +1953,7 @@ ], "model.layers.1.self_attn": [ { - "accuracy": 0.9072708263993263, + "accuracy": 0.9073722139000893, "total_bits": 30308928, "q_proj": { "group_size": { @@ -2017,7 +2017,7 @@ } }, { - "accuracy": 0.9114431366324425, + "accuracy": 0.91156817227602, "total_bits": 31455808, "q_proj": { "group_size": { @@ -2081,7 +2081,7 @@ } }, { - "accuracy": 0.9211425334215164, + "accuracy": 0.92121671885252, "total_bits": 33412832, "q_proj": { "group_size": { @@ -2145,7 +2145,7 @@ } }, { - "accuracy": 0.940546628087759, + "accuracy": 0.9407354593276978, "total_bits": 37983200, "q_proj": { "group_size": { @@ -2209,7 +2209,7 @@ } }, { - "accuracy": 0.9535201378166676, + "accuracy": 0.9535512514412403, "total_bits": 44838176, "q_proj": { "group_size": { @@ -2273,7 +2273,7 @@ } }, { - "accuracy": 0.9539150819182396, + "accuracy": 0.953955851495266, "total_bits": 44912768, "q_proj": { "group_size": { @@ -2337,7 +2337,7 @@ } }, { - "accuracy": 0.9746329039335251, + "accuracy": 0.9746336918324232, "total_bits": 57355552, "q_proj": { "group_size": { @@ -2389,7 +2389,7 @@ } }, { - "accuracy": 0.9750404637306929, + "accuracy": 0.9750551991164684, "total_bits": 57430144, "q_proj": { "group_size": { @@ -2441,7 +2441,7 @@ } }, { - "accuracy": 0.976738478988409, + "accuracy": 0.976717684417963, "total_bits": 57950464, "q_proj": { "group_size": { @@ -2493,7 +2493,7 @@ } }, { - "accuracy": 0.9780394490808249, + "accuracy": 0.9780480619519949, "total_bits": 58692736, "q_proj": { "group_size": { @@ -2545,7 +2545,7 @@ } }, { - "accuracy": 0.9763826839625835, + "accuracy": 0.9763856939971447, "total_bits": 59068544, "q_proj": { "group_size": { @@ -2609,7 +2609,7 @@ } }, { - "accuracy": 0.9780621360987425, + "accuracy": 0.9780414216220379, "total_bits": 59588864, "q_proj": { "group_size": { @@ -2673,7 +2673,7 @@ } }, { - "accuracy": 0.9801010321825743, + "accuracy": 0.9800688195973635, "total_bits": 61536832, "q_proj": { "group_size": { @@ -2734,7 +2734,7 @@ } }, { - "accuracy": 0.981600085273385, + "accuracy": 0.9815671853721142, "total_bits": 62354560, "q_proj": { "group_size": { @@ -2795,7 +2795,7 @@ } }, { - "accuracy": 0.9890466667711735, + "accuracy": 0.9890200421214104, "total_bits": 75246880, "q_proj": { "group_size": { @@ -2856,7 +2856,7 @@ } }, { - "accuracy": 0.9907460613176227, + "accuracy": 0.9907318344339728, "total_bits": 76510336, "q_proj": { "group_size": { @@ -2917,7 +2917,7 @@ } }, { - "accuracy": 0.9933391944505274, + "accuracy": 0.9933400782756507, "total_bits": 85667104, "q_proj": { "group_size": { @@ -2969,7 +2969,7 @@ } }, { - "accuracy": 0.9952039420604706, + "accuracy": 0.9951940830796957, "total_bits": 91722880, "q_proj": { "group_size": { @@ -3021,7 +3021,7 @@ } }, { - "accuracy": 0.9982604767428711, + "accuracy": 0.9982599728973582, "total_bits": 113978656, "q_proj": { "group_size": { @@ -3075,7 +3075,7 @@ ], "model.layers.1.mlp": [ { - "accuracy": 0.8869373872876167, + "accuracy": 0.887041375041008, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -3127,7 +3127,7 @@ } }, { - "accuracy": 0.891030453145504, + "accuracy": 0.8909792527556419, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -3179,7 +3179,7 @@ } }, { - "accuracy": 0.9027609676122665, + "accuracy": 0.9027382656931877, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -3228,7 +3228,7 @@ } }, { - "accuracy": 0.9062561988830566, + "accuracy": 0.9062387868762016, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -3277,7 +3277,7 @@ } }, { - "accuracy": 0.9437772929668427, + "accuracy": 0.943806029856205, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -3329,7 +3329,7 @@ } }, { - "accuracy": 0.9482904970645905, + "accuracy": 0.9483064226806164, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -3381,7 +3381,7 @@ } }, { - "accuracy": 0.952780719846487, + "accuracy": 0.9527781158685684, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -3430,7 +3430,7 @@ } }, { - "accuracy": 0.9698743838816881, + "accuracy": 0.9698769953101873, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -3473,7 +3473,7 @@ } }, { - "accuracy": 0.9721903912723064, + "accuracy": 0.9722175262868404, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -3516,7 +3516,7 @@ } }, { - "accuracy": 0.9712930209934711, + "accuracy": 0.9713117778301239, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -3568,7 +3568,7 @@ } }, { - "accuracy": 0.9747709520161152, + "accuracy": 0.9747742302715778, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -3620,7 +3620,7 @@ } }, { - "accuracy": 0.9853192111477256, + "accuracy": 0.9853277914226055, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -3672,7 +3672,7 @@ } }, { - "accuracy": 0.987357254140079, + "accuracy": 0.9873669799417257, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -3724,7 +3724,7 @@ } }, { - "accuracy": 0.9915956920012832, + "accuracy": 0.9915947569534183, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -3767,7 +3767,7 @@ } }, { - "accuracy": 0.9925124426372349, + "accuracy": 0.9925175970420241, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -3816,7 +3816,7 @@ } }, { - "accuracy": 0.9935486940667033, + "accuracy": 0.9935514670796692, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -3862,7 +3862,7 @@ } }, { - "accuracy": 0.9976398337166756, + "accuracy": 0.9976405743509531, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -3904,7 +3904,7 @@ ], "model.layers.2.self_attn": [ { - "accuracy": 0.8655510395765305, + "accuracy": 0.8655463457107544, "total_bits": 30308928, "q_proj": { "group_size": { @@ -3968,7 +3968,7 @@ } }, { - "accuracy": 0.8742070347070694, + "accuracy": 0.874164268374443, "total_bits": 31455808, "q_proj": { "group_size": { @@ -4032,7 +4032,7 @@ } }, { - "accuracy": 0.8933860436081886, + "accuracy": 0.8934662640094757, "total_bits": 33412832, "q_proj": { "group_size": { @@ -4096,7 +4096,7 @@ } }, { - "accuracy": 0.9261169582605362, + "accuracy": 0.926240935921669, "total_bits": 37983200, "q_proj": { "group_size": { @@ -4160,7 +4160,7 @@ } }, { - "accuracy": 0.933709405362606, + "accuracy": 0.9337229505181313, "total_bits": 44838176, "q_proj": { "group_size": { @@ -4224,7 +4224,7 @@ } }, { - "accuracy": 0.9347171932458878, + "accuracy": 0.9346923530101776, "total_bits": 44912768, "q_proj": { "group_size": { @@ -4288,7 +4288,7 @@ } }, { - "accuracy": 0.9634591825306416, + "accuracy": 0.9634601436555386, "total_bits": 57355552, "q_proj": { "group_size": { @@ -4340,7 +4340,7 @@ } }, { - "accuracy": 0.9642514251172543, + "accuracy": 0.9642136543989182, "total_bits": 57430144, "q_proj": { "group_size": { @@ -4392,7 +4392,7 @@ } }, { - "accuracy": 0.9666738845407963, + "accuracy": 0.9666857272386551, "total_bits": 57950464, "q_proj": { "group_size": { @@ -4444,7 +4444,7 @@ } }, { - "accuracy": 0.9683135040104389, + "accuracy": 0.9683275371789932, "total_bits": 58692736, "q_proj": { "group_size": { @@ -4496,7 +4496,7 @@ } }, { - "accuracy": 0.966770775616169, + "accuracy": 0.9667858779430389, "total_bits": 59068544, "q_proj": { "group_size": { @@ -4560,7 +4560,7 @@ } }, { - "accuracy": 0.9689657036215067, + "accuracy": 0.9689403381198645, "total_bits": 59588864, "q_proj": { "group_size": { @@ -4624,7 +4624,7 @@ } }, { - "accuracy": 0.9732588063925505, + "accuracy": 0.9732514582574368, "total_bits": 61536832, "q_proj": { "group_size": { @@ -4685,7 +4685,7 @@ } }, { - "accuracy": 0.9749700985848904, + "accuracy": 0.9749796185642481, "total_bits": 62354560, "q_proj": { "group_size": { @@ -4746,7 +4746,7 @@ } }, { - "accuracy": 0.985321925021708, + "accuracy": 0.9853210505098104, "total_bits": 75246880, "q_proj": { "group_size": { @@ -4807,7 +4807,7 @@ } }, { - "accuracy": 0.9874144792556763, + "accuracy": 0.9874111460521817, "total_bits": 76510336, "q_proj": { "group_size": { @@ -4868,7 +4868,7 @@ } }, { - "accuracy": 0.9904179144650698, + "accuracy": 0.9904154865071177, "total_bits": 85667104, "q_proj": { "group_size": { @@ -4920,7 +4920,7 @@ } }, { - "accuracy": 0.9935114970430732, + "accuracy": 0.9935085913166404, "total_bits": 91722880, "q_proj": { "group_size": { @@ -4972,7 +4972,7 @@ } }, { - "accuracy": 0.9975107992067933, + "accuracy": 0.9975118087604642, "total_bits": 113978656, "q_proj": { "group_size": { @@ -5026,7 +5026,7 @@ ], "model.layers.2.mlp": [ { - "accuracy": 0.8512221872806549, + "accuracy": 0.851137638092041, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -5078,7 +5078,7 @@ } }, { - "accuracy": 0.8556944578886032, + "accuracy": 0.8557578176259995, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -5130,7 +5130,7 @@ } }, { - "accuracy": 0.8742920905351639, + "accuracy": 0.8742973953485489, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -5179,7 +5179,7 @@ } }, { - "accuracy": 0.8798726350069046, + "accuracy": 0.8798764571547508, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -5228,7 +5228,7 @@ } }, { - "accuracy": 0.9260013550519943, + "accuracy": 0.9259674102067947, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -5280,7 +5280,7 @@ } }, { - "accuracy": 0.9317035898566246, + "accuracy": 0.9316787645220757, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -5332,7 +5332,7 @@ } }, { - "accuracy": 0.9391488991677761, + "accuracy": 0.9391309916973114, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -5381,7 +5381,7 @@ } }, { - "accuracy": 0.9611527696251869, + "accuracy": 0.9611707739531994, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -5424,7 +5424,7 @@ } }, { - "accuracy": 0.9643270559608936, + "accuracy": 0.964320108294487, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -5467,7 +5467,7 @@ } }, { - "accuracy": 0.9622061587870121, + "accuracy": 0.962238498032093, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -5519,7 +5519,7 @@ } }, { - "accuracy": 0.966707780957222, + "accuracy": 0.9667122215032578, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -5571,7 +5571,7 @@ } }, { - "accuracy": 0.9806768018752337, + "accuracy": 0.9806710071861744, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -5623,7 +5623,7 @@ } }, { - "accuracy": 0.9833411537110806, + "accuracy": 0.9833368640393019, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -5675,7 +5675,7 @@ } }, { - "accuracy": 0.9893169151619077, + "accuracy": 0.98932437133044, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -5718,7 +5718,7 @@ } }, { - "accuracy": 0.9900901559740305, + "accuracy": 0.990092589519918, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -5767,7 +5767,7 @@ } }, { - "accuracy": 0.9917803723365068, + "accuracy": 0.9917868180200458, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -5813,7 +5813,7 @@ } }, { - "accuracy": 0.997135940939188, + "accuracy": 0.9971363425720483, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -5855,7 +5855,7 @@ ], "model.layers.3.self_attn": [ { - "accuracy": 0.8845625817775726, + "accuracy": 0.8846139460802078, "total_bits": 30308928, "q_proj": { "group_size": { @@ -5919,7 +5919,7 @@ } }, { - "accuracy": 0.8901228457689285, + "accuracy": 0.8900518491864204, "total_bits": 31455808, "q_proj": { "group_size": { @@ -5983,7 +5983,7 @@ } }, { - "accuracy": 0.9061683937907219, + "accuracy": 0.9061124995350838, "total_bits": 33412832, "q_proj": { "group_size": { @@ -6047,7 +6047,7 @@ } }, { - "accuracy": 0.9347674697637558, + "accuracy": 0.9347528889775276, "total_bits": 37983200, "q_proj": { "group_size": { @@ -6111,7 +6111,7 @@ } }, { - "accuracy": 0.9422827735543251, + "accuracy": 0.9422792084515095, "total_bits": 44838176, "q_proj": { "group_size": { @@ -6175,7 +6175,7 @@ } }, { - "accuracy": 0.943145889788866, + "accuracy": 0.9431353025138378, "total_bits": 44912768, "q_proj": { "group_size": { @@ -6239,7 +6239,7 @@ } }, { - "accuracy": 0.9687017947435379, + "accuracy": 0.9687346704304218, "total_bits": 57355552, "q_proj": { "group_size": { @@ -6291,7 +6291,7 @@ } }, { - "accuracy": 0.969449121505022, + "accuracy": 0.9694707691669464, "total_bits": 57430144, "q_proj": { "group_size": { @@ -6343,7 +6343,7 @@ } }, { - "accuracy": 0.9714747574180365, + "accuracy": 0.9715053513646126, "total_bits": 57950464, "q_proj": { "group_size": { @@ -6395,7 +6395,7 @@ } }, { - "accuracy": 0.9727979246526957, + "accuracy": 0.9728143084794283, "total_bits": 58692736, "q_proj": { "group_size": { @@ -6447,7 +6447,7 @@ } }, { - "accuracy": 0.9711320530623198, + "accuracy": 0.9711146093904972, "total_bits": 59068544, "q_proj": { "group_size": { @@ -6511,7 +6511,7 @@ } }, { - "accuracy": 0.9730110038071871, + "accuracy": 0.9730188455432653, "total_bits": 59588864, "q_proj": { "group_size": { @@ -6575,7 +6575,7 @@ } }, { - "accuracy": 0.9766805954277515, + "accuracy": 0.9766947887837887, "total_bits": 61536832, "q_proj": { "group_size": { @@ -6636,7 +6636,7 @@ } }, { - "accuracy": 0.9783372972160578, + "accuracy": 0.9783317521214485, "total_bits": 62354560, "q_proj": { "group_size": { @@ -6697,7 +6697,7 @@ } }, { - "accuracy": 0.9873075932264328, + "accuracy": 0.9873018255457282, "total_bits": 75246880, "q_proj": { "group_size": { @@ -6758,7 +6758,7 @@ } }, { - "accuracy": 0.9891520766541362, + "accuracy": 0.9891604781150818, "total_bits": 76510336, "q_proj": { "group_size": { @@ -6819,7 +6819,7 @@ } }, { - "accuracy": 0.9918732857331634, + "accuracy": 0.9918712023645639, "total_bits": 85667104, "q_proj": { "group_size": { @@ -6871,7 +6871,7 @@ } }, { - "accuracy": 0.994609275367111, + "accuracy": 0.9946046634577215, "total_bits": 91722880, "q_proj": { "group_size": { @@ -6923,7 +6923,7 @@ } }, { - "accuracy": 0.9978871110361069, + "accuracy": 0.9978904649615288, "total_bits": 113978656, "q_proj": { "group_size": { @@ -6977,7 +6977,7 @@ ], "model.layers.3.mlp": [ { - "accuracy": 0.8215165734291077, + "accuracy": 0.8215235769748688, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -7029,7 +7029,7 @@ } }, { - "accuracy": 0.8267818093299866, + "accuracy": 0.8267826288938522, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -7081,7 +7081,7 @@ } }, { - "accuracy": 0.8501903116703033, + "accuracy": 0.8501260429620743, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -7130,7 +7130,7 @@ } }, { - "accuracy": 0.8572592884302139, + "accuracy": 0.8572545945644379, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -7179,7 +7179,7 @@ } }, { - "accuracy": 0.9100245088338852, + "accuracy": 0.9100974574685097, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -7231,7 +7231,7 @@ } }, { - "accuracy": 0.9172510206699371, + "accuracy": 0.9171888679265976, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -7283,7 +7283,7 @@ } }, { - "accuracy": 0.9271175712347031, + "accuracy": 0.92706498503685, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -7332,7 +7332,7 @@ } }, { - "accuracy": 0.953081201761961, + "accuracy": 0.9530698768794537, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -7375,7 +7375,7 @@ } }, { - "accuracy": 0.9570418335497379, + "accuracy": 0.957017607986927, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -7418,7 +7418,7 @@ } }, { - "accuracy": 0.9540456458926201, + "accuracy": 0.9540662579238415, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -7470,7 +7470,7 @@ } }, { - "accuracy": 0.9596328996121883, + "accuracy": 0.9596161395311356, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -7522,7 +7522,7 @@ } }, { - "accuracy": 0.9765274990350008, + "accuracy": 0.9765002690255642, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -7574,7 +7574,7 @@ } }, { - "accuracy": 0.9798263423144817, + "accuracy": 0.979815537109971, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -7626,7 +7626,7 @@ } }, { - "accuracy": 0.9871442606672645, + "accuracy": 0.987149802967906, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -7669,7 +7669,7 @@ } }, { - "accuracy": 0.9879334066063166, + "accuracy": 0.9879286577925086, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -7718,7 +7718,7 @@ } }, { - "accuracy": 0.9901348492130637, + "accuracy": 0.9901374354958534, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -7764,7 +7764,7 @@ } }, { - "accuracy": 0.9965592622756958, + "accuracy": 0.9965564699377865, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -7806,7 +7806,7 @@ ], "model.layers.4.self_attn": [ { - "accuracy": 0.8741211891174316, + "accuracy": 0.8741328567266464, "total_bits": 30308928, "q_proj": { "group_size": { @@ -7870,7 +7870,7 @@ } }, { - "accuracy": 0.8816789761185646, + "accuracy": 0.8816472738981247, "total_bits": 31455808, "q_proj": { "group_size": { @@ -7934,7 +7934,7 @@ } }, { - "accuracy": 0.9003885015845299, + "accuracy": 0.900344654917717, "total_bits": 33412832, "q_proj": { "group_size": { @@ -7998,7 +7998,7 @@ } }, { - "accuracy": 0.9307805150747299, + "accuracy": 0.9307383000850677, "total_bits": 37983200, "q_proj": { "group_size": { @@ -8062,7 +8062,7 @@ } }, { - "accuracy": 0.9374072849750519, + "accuracy": 0.9374280199408531, "total_bits": 44838176, "q_proj": { "group_size": { @@ -8126,7 +8126,7 @@ } }, { - "accuracy": 0.9384637139737606, + "accuracy": 0.9384343661367893, "total_bits": 44912768, "q_proj": { "group_size": { @@ -8190,7 +8190,7 @@ } }, { - "accuracy": 0.9653116390109062, + "accuracy": 0.9653166085481644, "total_bits": 57355552, "q_proj": { "group_size": { @@ -8242,7 +8242,7 @@ } }, { - "accuracy": 0.9661814905703068, + "accuracy": 0.9661872051656246, "total_bits": 57430144, "q_proj": { "group_size": { @@ -8294,7 +8294,7 @@ } }, { - "accuracy": 0.9684111662209034, + "accuracy": 0.968420397490263, "total_bits": 57950464, "q_proj": { "group_size": { @@ -8346,7 +8346,7 @@ } }, { - "accuracy": 0.969778697937727, + "accuracy": 0.9697968028485775, "total_bits": 58692736, "q_proj": { "group_size": { @@ -8398,7 +8398,7 @@ } }, { - "accuracy": 0.9687559828162193, + "accuracy": 0.9687640070915222, "total_bits": 59068544, "q_proj": { "group_size": { @@ -8462,7 +8462,7 @@ } }, { - "accuracy": 0.9707713481038809, + "accuracy": 0.9707700200378895, "total_bits": 59588864, "q_proj": { "group_size": { @@ -8526,7 +8526,7 @@ } }, { - "accuracy": 0.975062221288681, + "accuracy": 0.9750581458210945, "total_bits": 61536832, "q_proj": { "group_size": { @@ -8587,7 +8587,7 @@ } }, { - "accuracy": 0.9766068141907454, + "accuracy": 0.9766066540032625, "total_bits": 62354560, "q_proj": { "group_size": { @@ -8648,7 +8648,7 @@ } }, { - "accuracy": 0.986367778852582, + "accuracy": 0.9863643515855074, "total_bits": 75246880, "q_proj": { "group_size": { @@ -8709,7 +8709,7 @@ } }, { - "accuracy": 0.9882961716502905, + "accuracy": 0.9882954657077789, "total_bits": 76510336, "q_proj": { "group_size": { @@ -8770,7 +8770,7 @@ } }, { - "accuracy": 0.9909742707386613, + "accuracy": 0.9909781841561198, "total_bits": 85667104, "q_proj": { "group_size": { @@ -8822,7 +8822,7 @@ } }, { - "accuracy": 0.9939752677455544, + "accuracy": 0.9939774051308632, "total_bits": 91722880, "q_proj": { "group_size": { @@ -8874,7 +8874,7 @@ } }, { - "accuracy": 0.9976646257564425, + "accuracy": 0.9976642981637269, "total_bits": 113978656, "q_proj": { "group_size": { @@ -8928,7 +8928,7 @@ ], "model.layers.4.mlp": [ { - "accuracy": 0.8349805325269699, + "accuracy": 0.8348695486783981, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -8980,7 +8980,7 @@ } }, { - "accuracy": 0.8401748985052109, + "accuracy": 0.8400902897119522, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -9032,7 +9032,7 @@ } }, { - "accuracy": 0.8632630258798599, + "accuracy": 0.863140657544136, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -9081,7 +9081,7 @@ } }, { - "accuracy": 0.8702157586812973, + "accuracy": 0.870071679353714, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -9130,7 +9130,7 @@ } }, { - "accuracy": 0.9168255552649498, + "accuracy": 0.9167942702770233, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -9182,7 +9182,7 @@ } }, { - "accuracy": 0.9239948987960815, + "accuracy": 0.9239982962608337, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -9234,7 +9234,7 @@ } }, { - "accuracy": 0.933478482067585, + "accuracy": 0.9335077553987503, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -9283,7 +9283,7 @@ } }, { - "accuracy": 0.9563411399722099, + "accuracy": 0.9563678838312626, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -9326,7 +9326,7 @@ } }, { - "accuracy": 0.960248950868845, + "accuracy": 0.960248876363039, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -9369,7 +9369,7 @@ } }, { - "accuracy": 0.9574717655777931, + "accuracy": 0.9574565887451172, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -9421,7 +9421,7 @@ } }, { - "accuracy": 0.9629541300237179, + "accuracy": 0.9629608169198036, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -9473,7 +9473,7 @@ } }, { - "accuracy": 0.9782424960285425, + "accuracy": 0.9782421588897705, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -9525,7 +9525,7 @@ } }, { - "accuracy": 0.9814701918512583, + "accuracy": 0.9814739804714918, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -9577,7 +9577,7 @@ } }, { - "accuracy": 0.9879318736493587, + "accuracy": 0.9879294466227293, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -9620,7 +9620,7 @@ } }, { - "accuracy": 0.9888098947703838, + "accuracy": 0.988809896633029, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -9669,7 +9669,7 @@ } }, { - "accuracy": 0.9909165976569057, + "accuracy": 0.9909201338887215, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -9715,7 +9715,7 @@ } }, { - "accuracy": 0.9967805305495858, + "accuracy": 0.9967838707379997, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -9757,7 +9757,7 @@ ], "model.layers.5.self_attn": [ { - "accuracy": 0.8810675218701363, + "accuracy": 0.8811230733990669, "total_bits": 30308928, "q_proj": { "group_size": { @@ -9821,7 +9821,7 @@ } }, { - "accuracy": 0.8886952474713326, + "accuracy": 0.8887623399496078, "total_bits": 31455808, "q_proj": { "group_size": { @@ -9885,7 +9885,7 @@ } }, { - "accuracy": 0.9045190438628197, + "accuracy": 0.904540665447712, "total_bits": 33412832, "q_proj": { "group_size": { @@ -9949,7 +9949,7 @@ } }, { - "accuracy": 0.9352795407176018, + "accuracy": 0.9353184551000595, "total_bits": 37983200, "q_proj": { "group_size": { @@ -10013,7 +10013,7 @@ } }, { - "accuracy": 0.9399208948016167, + "accuracy": 0.9399426616728306, "total_bits": 44838176, "q_proj": { "group_size": { @@ -10077,7 +10077,7 @@ } }, { - "accuracy": 0.9410183131694794, + "accuracy": 0.9409984610974789, "total_bits": 44912768, "q_proj": { "group_size": { @@ -10141,7 +10141,7 @@ } }, { - "accuracy": 0.9663527235388756, + "accuracy": 0.9663707427680492, "total_bits": 57355552, "q_proj": { "group_size": { @@ -10193,7 +10193,7 @@ } }, { - "accuracy": 0.967367872595787, + "accuracy": 0.9673691280186176, "total_bits": 57430144, "q_proj": { "group_size": { @@ -10245,7 +10245,7 @@ } }, { - "accuracy": 0.9696921166032553, + "accuracy": 0.9697138108313084, "total_bits": 57950464, "q_proj": { "group_size": { @@ -10297,7 +10297,7 @@ } }, { - "accuracy": 0.9711299203336239, + "accuracy": 0.9711332526057959, "total_bits": 58692736, "q_proj": { "group_size": { @@ -10349,7 +10349,7 @@ } }, { - "accuracy": 0.9700027704238892, + "accuracy": 0.9700067639350891, "total_bits": 59068544, "q_proj": { "group_size": { @@ -10413,7 +10413,7 @@ } }, { - "accuracy": 0.9720256011933088, + "accuracy": 0.9720334056764841, "total_bits": 59588864, "q_proj": { "group_size": { @@ -10477,7 +10477,7 @@ } }, { - "accuracy": 0.9756313841789961, + "accuracy": 0.9756301864981651, "total_bits": 61536832, "q_proj": { "group_size": { @@ -10538,7 +10538,7 @@ } }, { - "accuracy": 0.9773869439959526, + "accuracy": 0.9774045348167419, "total_bits": 62354560, "q_proj": { "group_size": { @@ -10599,7 +10599,7 @@ } }, { - "accuracy": 0.9866560539230704, + "accuracy": 0.9866563268005848, "total_bits": 75246880, "q_proj": { "group_size": { @@ -10660,7 +10660,7 @@ } }, { - "accuracy": 0.9886289816349745, + "accuracy": 0.9886361388489604, "total_bits": 76510336, "q_proj": { "group_size": { @@ -10721,7 +10721,7 @@ } }, { - "accuracy": 0.9912481410428882, + "accuracy": 0.9912527557462454, "total_bits": 85667104, "q_proj": { "group_size": { @@ -10773,7 +10773,7 @@ } }, { - "accuracy": 0.9942180886864662, + "accuracy": 0.9942179205827415, "total_bits": 91722880, "q_proj": { "group_size": { @@ -10825,7 +10825,7 @@ } }, { - "accuracy": 0.9977244043257087, + "accuracy": 0.9977258909493685, "total_bits": 113978656, "q_proj": { "group_size": { @@ -10879,7 +10879,7 @@ ], "model.layers.5.mlp": [ { - "accuracy": 0.8648017942905426, + "accuracy": 0.8648174107074738, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -10931,7 +10931,7 @@ } }, { - "accuracy": 0.8694759905338287, + "accuracy": 0.8695154935121536, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -10983,7 +10983,7 @@ } }, { - "accuracy": 0.8880036026239395, + "accuracy": 0.888036236166954, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -11032,7 +11032,7 @@ } }, { - "accuracy": 0.8937847763299942, + "accuracy": 0.8938298374414444, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -11081,7 +11081,7 @@ } }, { - "accuracy": 0.9318181350827217, + "accuracy": 0.931794673204422, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -11133,7 +11133,7 @@ } }, { - "accuracy": 0.9377684108912945, + "accuracy": 0.9377873539924622, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -11185,7 +11185,7 @@ } }, { - "accuracy": 0.9455864503979683, + "accuracy": 0.9456103965640068, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -11234,7 +11234,7 @@ } }, { - "accuracy": 0.9638955146074295, + "accuracy": 0.963904220610857, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -11277,7 +11277,7 @@ } }, { - "accuracy": 0.9671880304813385, + "accuracy": 0.9671907536685467, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -11320,7 +11320,7 @@ } }, { - "accuracy": 0.9650383368134499, + "accuracy": 0.9650482423603535, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -11372,7 +11372,7 @@ } }, { - "accuracy": 0.9695848729461432, + "accuracy": 0.9695756994187832, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -11424,7 +11424,7 @@ } }, { - "accuracy": 0.9821004029363394, + "accuracy": 0.982111718505621, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -11476,7 +11476,7 @@ } }, { - "accuracy": 0.9847700102254748, + "accuracy": 0.984773620031774, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -11528,7 +11528,7 @@ } }, { - "accuracy": 0.990053579211235, + "accuracy": 0.9900560518726707, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -11571,7 +11571,7 @@ } }, { - "accuracy": 0.9907908504828811, + "accuracy": 0.990793714299798, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -11620,7 +11620,7 @@ } }, { - "accuracy": 0.992512381169945, + "accuracy": 0.992513523902744, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -11666,7 +11666,7 @@ } }, { - "accuracy": 0.9973284220322967, + "accuracy": 0.997328422497958, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -11708,7 +11708,7 @@ ], "model.layers.6.self_attn": [ { - "accuracy": 0.8890528008341789, + "accuracy": 0.8890272378921509, "total_bits": 30308928, "q_proj": { "group_size": { @@ -11772,7 +11772,7 @@ } }, { - "accuracy": 0.8953105807304382, + "accuracy": 0.8953426331281662, "total_bits": 31455808, "q_proj": { "group_size": { @@ -11836,7 +11836,7 @@ } }, { - "accuracy": 0.9070602431893349, + "accuracy": 0.9071049243211746, "total_bits": 33412832, "q_proj": { "group_size": { @@ -11900,7 +11900,7 @@ } }, { - "accuracy": 0.9330306127667427, + "accuracy": 0.9330921769142151, "total_bits": 37983200, "q_proj": { "group_size": { @@ -11964,7 +11964,7 @@ } }, { - "accuracy": 0.9448112137615681, + "accuracy": 0.9447792358696461, "total_bits": 44838176, "q_proj": { "group_size": { @@ -12028,7 +12028,7 @@ } }, { - "accuracy": 0.9454820677638054, + "accuracy": 0.9455131739377975, "total_bits": 44912768, "q_proj": { "group_size": { @@ -12092,7 +12092,7 @@ } }, { - "accuracy": 0.9695140719413757, + "accuracy": 0.969525508582592, "total_bits": 57355552, "q_proj": { "group_size": { @@ -12144,7 +12144,7 @@ } }, { - "accuracy": 0.9701519533991814, + "accuracy": 0.9701721202582121, "total_bits": 57430144, "q_proj": { "group_size": { @@ -12196,7 +12196,7 @@ } }, { - "accuracy": 0.972233023494482, + "accuracy": 0.9722142405807972, "total_bits": 57950464, "q_proj": { "group_size": { @@ -12248,7 +12248,7 @@ } }, { - "accuracy": 0.9737806580960751, + "accuracy": 0.9737892020493746, "total_bits": 58692736, "q_proj": { "group_size": { @@ -12300,7 +12300,7 @@ } }, { - "accuracy": 0.9722792375832796, + "accuracy": 0.9722646549344063, "total_bits": 59068544, "q_proj": { "group_size": { @@ -12364,7 +12364,7 @@ } }, { - "accuracy": 0.9740959145128727, + "accuracy": 0.9741003829985857, "total_bits": 59588864, "q_proj": { "group_size": { @@ -12428,7 +12428,7 @@ } }, { - "accuracy": 0.9769062623381615, + "accuracy": 0.9769276678562164, "total_bits": 61536832, "q_proj": { "group_size": { @@ -12489,7 +12489,7 @@ } }, { - "accuracy": 0.9786777105182409, + "accuracy": 0.9786724224686623, "total_bits": 62354560, "q_proj": { "group_size": { @@ -12550,7 +12550,7 @@ } }, { - "accuracy": 0.987405676394701, + "accuracy": 0.9874210571870208, "total_bits": 75246880, "q_proj": { "group_size": { @@ -12611,7 +12611,7 @@ } }, { - "accuracy": 0.9893234008923173, + "accuracy": 0.9893340524286032, "total_bits": 76510336, "q_proj": { "group_size": { @@ -12672,7 +12672,7 @@ } }, { - "accuracy": 0.9920955216512084, + "accuracy": 0.9920956883579493, "total_bits": 85667104, "q_proj": { "group_size": { @@ -12724,7 +12724,7 @@ } }, { - "accuracy": 0.9945340054109693, + "accuracy": 0.9945395006798208, "total_bits": 91722880, "q_proj": { "group_size": { @@ -12776,7 +12776,7 @@ } }, { - "accuracy": 0.9979429002851248, + "accuracy": 0.9979436090216041, "total_bits": 113978656, "q_proj": { "group_size": { @@ -12830,7 +12830,7 @@ ], "model.layers.6.mlp": [ { - "accuracy": 0.8601508587598801, + "accuracy": 0.860052615404129, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -12882,7 +12882,7 @@ } }, { - "accuracy": 0.8646412193775177, + "accuracy": 0.864541083574295, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -12934,7 +12934,7 @@ } }, { - "accuracy": 0.883434846997261, + "accuracy": 0.8833191320300102, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -12983,7 +12983,7 @@ } }, { - "accuracy": 0.8892851322889328, + "accuracy": 0.8891885280609131, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -13032,7 +13032,7 @@ } }, { - "accuracy": 0.9288300052285194, + "accuracy": 0.928782157599926, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -13084,7 +13084,7 @@ } }, { - "accuracy": 0.9350205287337303, + "accuracy": 0.9350093528628349, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -13136,7 +13136,7 @@ } }, { - "accuracy": 0.942976824939251, + "accuracy": 0.9429488480091095, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -13185,7 +13185,7 @@ } }, { - "accuracy": 0.9626522436738014, + "accuracy": 0.9626510068774223, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -13228,7 +13228,7 @@ } }, { - "accuracy": 0.9660748802125454, + "accuracy": 0.9660806246101856, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -13271,7 +13271,7 @@ } }, { - "accuracy": 0.9634818024933338, + "accuracy": 0.9634533040225506, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -13323,7 +13323,7 @@ } }, { - "accuracy": 0.9682779163122177, + "accuracy": 0.9682538993656635, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -13375,7 +13375,7 @@ } }, { - "accuracy": 0.9812953136861324, + "accuracy": 0.9812956769019365, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -13427,7 +13427,7 @@ } }, { - "accuracy": 0.9841256625950336, + "accuracy": 0.9841305427253246, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -13479,7 +13479,7 @@ } }, { - "accuracy": 0.9896932039409876, + "accuracy": 0.9896902348846197, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -13522,7 +13522,7 @@ } }, { - "accuracy": 0.9903836958110332, + "accuracy": 0.9903781078755856, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -13571,7 +13571,7 @@ } }, { - "accuracy": 0.9921105708926916, + "accuracy": 0.9921116251498461, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -13617,7 +13617,7 @@ } }, { - "accuracy": 0.9972038897685707, + "accuracy": 0.9972023773007095, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -13659,7 +13659,7 @@ ], "model.layers.7.self_attn": [ { - "accuracy": 0.8790968656539917, + "accuracy": 0.879193864762783, "total_bits": 30308928, "q_proj": { "group_size": { @@ -13723,7 +13723,7 @@ } }, { - "accuracy": 0.8855356946587563, + "accuracy": 0.885476142168045, "total_bits": 31455808, "q_proj": { "group_size": { @@ -13787,7 +13787,7 @@ } }, { - "accuracy": 0.9016124606132507, + "accuracy": 0.9015669599175453, "total_bits": 33412832, "q_proj": { "group_size": { @@ -13851,7 +13851,7 @@ } }, { - "accuracy": 0.9258091002702713, + "accuracy": 0.9257946684956551, "total_bits": 37983200, "q_proj": { "group_size": { @@ -13915,7 +13915,7 @@ } }, { - "accuracy": 0.9374926388263702, + "accuracy": 0.9375413469970226, "total_bits": 44838176, "q_proj": { "group_size": { @@ -13979,7 +13979,7 @@ } }, { - "accuracy": 0.9384367614984512, + "accuracy": 0.9384141005575657, "total_bits": 44912768, "q_proj": { "group_size": { @@ -14043,7 +14043,7 @@ } }, { - "accuracy": 0.9647591635584831, + "accuracy": 0.9647888205945492, "total_bits": 57355552, "q_proj": { "group_size": { @@ -14095,7 +14095,7 @@ } }, { - "accuracy": 0.965646505355835, + "accuracy": 0.9656654335558414, "total_bits": 57430144, "q_proj": { "group_size": { @@ -14147,7 +14147,7 @@ } }, { - "accuracy": 0.9682505577802658, + "accuracy": 0.968209270387888, "total_bits": 57950464, "q_proj": { "group_size": { @@ -14199,7 +14199,7 @@ } }, { - "accuracy": 0.9700952749699354, + "accuracy": 0.9700861666351557, "total_bits": 58692736, "q_proj": { "group_size": { @@ -14251,7 +14251,7 @@ } }, { - "accuracy": 0.9685632511973381, + "accuracy": 0.9685384854674339, "total_bits": 59068544, "q_proj": { "group_size": { @@ -14315,7 +14315,7 @@ } }, { - "accuracy": 0.9708435665816069, + "accuracy": 0.9708608984947205, "total_bits": 59588864, "q_proj": { "group_size": { @@ -14379,7 +14379,7 @@ } }, { - "accuracy": 0.9744916334748268, + "accuracy": 0.9745250958949327, "total_bits": 61536832, "q_proj": { "group_size": { @@ -14440,7 +14440,7 @@ } }, { - "accuracy": 0.976450975984335, + "accuracy": 0.9764064699411392, "total_bits": 62354560, "q_proj": { "group_size": { @@ -14501,7 +14501,7 @@ } }, { - "accuracy": 0.9858461897820234, + "accuracy": 0.9858506005257368, "total_bits": 75246880, "q_proj": { "group_size": { @@ -14562,7 +14562,7 @@ } }, { - "accuracy": 0.9881619503721595, + "accuracy": 0.9881491707637906, "total_bits": 76510336, "q_proj": { "group_size": { @@ -14623,7 +14623,7 @@ } }, { - "accuracy": 0.9907449893653393, + "accuracy": 0.9907346172258258, "total_bits": 85667104, "q_proj": { "group_size": { @@ -14675,7 +14675,7 @@ } }, { - "accuracy": 0.9937627115286887, + "accuracy": 0.9937750529497862, "total_bits": 91722880, "q_proj": { "group_size": { @@ -14727,7 +14727,7 @@ } }, { - "accuracy": 0.9975894596427679, + "accuracy": 0.997590501094237, "total_bits": 113978656, "q_proj": { "group_size": { @@ -14781,7 +14781,7 @@ ], "model.layers.7.mlp": [ { - "accuracy": 0.8853595852851868, + "accuracy": 0.8853648155927658, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -14833,7 +14833,7 @@ } }, { - "accuracy": 0.8890897259116173, + "accuracy": 0.8890001177787781, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -14885,7 +14885,7 @@ } }, { - "accuracy": 0.9030437618494034, + "accuracy": 0.9031427130103111, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -14934,7 +14934,7 @@ } }, { - "accuracy": 0.9074160978198051, + "accuracy": 0.9074975475668907, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -14983,7 +14983,7 @@ } }, { - "accuracy": 0.941648468375206, + "accuracy": 0.9416834153234959, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -15035,7 +15035,7 @@ } }, { - "accuracy": 0.9470459222793579, + "accuracy": 0.9470086395740509, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -15087,7 +15087,7 @@ } }, { - "accuracy": 0.9530564919114113, + "accuracy": 0.9530163891613483, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -15136,7 +15136,7 @@ } }, { - "accuracy": 0.9691261779516935, + "accuracy": 0.9691323079168797, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -15179,7 +15179,7 @@ } }, { - "accuracy": 0.9720455892384052, + "accuracy": 0.9720457717776299, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -15222,7 +15222,7 @@ } }, { - "accuracy": 0.9700704663991928, + "accuracy": 0.9700757917016745, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -15274,7 +15274,7 @@ } }, { - "accuracy": 0.9741590898483992, + "accuracy": 0.9741606414318085, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -15326,7 +15326,7 @@ } }, { - "accuracy": 0.9846790870651603, + "accuracy": 0.9846808835864067, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -15378,7 +15378,7 @@ } }, { - "accuracy": 0.9870709776878357, + "accuracy": 0.9870640141889453, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -15430,7 +15430,7 @@ } }, { - "accuracy": 0.9914329303428531, + "accuracy": 0.9914321266114712, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -15473,7 +15473,7 @@ } }, { - "accuracy": 0.9921443285420537, + "accuracy": 0.9921473953872919, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -15522,7 +15522,7 @@ } }, { - "accuracy": 0.9934511734172702, + "accuracy": 0.9934548917226493, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -15568,7 +15568,7 @@ } }, { - "accuracy": 0.997653994243592, + "accuracy": 0.9976542603690177, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -15610,7 +15610,7 @@ ], "model.layers.8.self_attn": [ { - "accuracy": 0.9262347742915154, + "accuracy": 0.9262180775403976, "total_bits": 30308928, "q_proj": { "group_size": { @@ -15674,7 +15674,7 @@ } }, { - "accuracy": 0.9292053729295731, + "accuracy": 0.9292021617293358, "total_bits": 31455808, "q_proj": { "group_size": { @@ -15738,7 +15738,7 @@ } }, { - "accuracy": 0.9359699636697769, + "accuracy": 0.9360192567110062, "total_bits": 33412832, "q_proj": { "group_size": { @@ -15802,7 +15802,7 @@ } }, { - "accuracy": 0.9522256478667259, + "accuracy": 0.9522725865244865, "total_bits": 37983200, "q_proj": { "group_size": { @@ -15866,7 +15866,7 @@ } }, { - "accuracy": 0.9621881134808064, + "accuracy": 0.9622119329869747, "total_bits": 44838176, "q_proj": { "group_size": { @@ -15930,7 +15930,7 @@ } }, { - "accuracy": 0.9625761918723583, + "accuracy": 0.9625555910170078, "total_bits": 44912768, "q_proj": { "group_size": { @@ -15994,7 +15994,7 @@ } }, { - "accuracy": 0.9790825769305229, + "accuracy": 0.9790912084281445, "total_bits": 57355552, "q_proj": { "group_size": { @@ -16046,7 +16046,7 @@ } }, { - "accuracy": 0.9794511869549751, + "accuracy": 0.9794544521719217, "total_bits": 57430144, "q_proj": { "group_size": { @@ -16098,7 +16098,7 @@ } }, { - "accuracy": 0.9807918332517147, + "accuracy": 0.9808075986802578, "total_bits": 57950464, "q_proj": { "group_size": { @@ -16150,7 +16150,7 @@ } }, { - "accuracy": 0.9819746408611536, + "accuracy": 0.9819814618676901, "total_bits": 58692736, "q_proj": { "group_size": { @@ -16202,7 +16202,7 @@ } }, { - "accuracy": 0.9808401670306921, + "accuracy": 0.980836171656847, "total_bits": 59068544, "q_proj": { "group_size": { @@ -16266,7 +16266,7 @@ } }, { - "accuracy": 0.9820886384695768, + "accuracy": 0.9820813406258821, "total_bits": 59588864, "q_proj": { "group_size": { @@ -16330,7 +16330,7 @@ } }, { - "accuracy": 0.9836654253304005, + "accuracy": 0.9836451895534992, "total_bits": 61536832, "q_proj": { "group_size": { @@ -16391,7 +16391,7 @@ } }, { - "accuracy": 0.984879620373249, + "accuracy": 0.9848749991506338, "total_bits": 62354560, "q_proj": { "group_size": { @@ -16452,7 +16452,7 @@ } }, { - "accuracy": 0.9910424621775746, + "accuracy": 0.9910451974719763, "total_bits": 75246880, "q_proj": { "group_size": { @@ -16513,7 +16513,7 @@ } }, { - "accuracy": 0.9924080655910075, + "accuracy": 0.9923991961404681, "total_bits": 76510336, "q_proj": { "group_size": { @@ -16574,7 +16574,7 @@ } }, { - "accuracy": 0.9945205664262176, + "accuracy": 0.9945177044719458, "total_bits": 85667104, "q_proj": { "group_size": { @@ -16626,7 +16626,7 @@ } }, { - "accuracy": 0.996047873981297, + "accuracy": 0.9960471182130277, "total_bits": 91722880, "q_proj": { "group_size": { @@ -16678,7 +16678,7 @@ } }, { - "accuracy": 0.998559134779498, + "accuracy": 0.9985587185947224, "total_bits": 113978656, "q_proj": { "group_size": { @@ -16732,7 +16732,7 @@ ], "model.layers.8.mlp": [ { - "accuracy": 0.8776565492153168, + "accuracy": 0.8774635121226311, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -16784,7 +16784,7 @@ } }, { - "accuracy": 0.8816645741462708, + "accuracy": 0.8816986083984375, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -16836,7 +16836,7 @@ } }, { - "accuracy": 0.8966159075498581, + "accuracy": 0.8964935094118118, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -16885,7 +16885,7 @@ } }, { - "accuracy": 0.9011181369423866, + "accuracy": 0.9010185077786446, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -16934,7 +16934,7 @@ } }, { - "accuracy": 0.9383117109537125, + "accuracy": 0.9382704459130764, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -16986,7 +16986,7 @@ } }, { - "accuracy": 0.9435076154768467, + "accuracy": 0.943505771458149, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -17038,7 +17038,7 @@ } }, { - "accuracy": 0.94978042319417, + "accuracy": 0.9497714042663574, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -17087,7 +17087,7 @@ } }, { - "accuracy": 0.9677746780216694, + "accuracy": 0.9677646122872829, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -17130,7 +17130,7 @@ } }, { - "accuracy": 0.9705556537955999, + "accuracy": 0.9705599583685398, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -17173,7 +17173,7 @@ } }, { - "accuracy": 0.9684275537729263, + "accuracy": 0.9684217162430286, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -17225,7 +17225,7 @@ } }, { - "accuracy": 0.972480921074748, + "accuracy": 0.9724891372025013, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -17277,7 +17277,7 @@ } }, { - "accuracy": 0.9838621038943529, + "accuracy": 0.9838705994188786, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -17329,7 +17329,7 @@ } }, { - "accuracy": 0.9862669911235571, + "accuracy": 0.9862625077366829, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -17381,7 +17381,7 @@ } }, { - "accuracy": 0.9911140948534012, + "accuracy": 0.9911155067384243, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -17424,7 +17424,7 @@ } }, { - "accuracy": 0.9916985612362623, + "accuracy": 0.9917009193450212, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -17473,7 +17473,7 @@ } }, { - "accuracy": 0.9930045073851943, + "accuracy": 0.9930097297765315, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -17519,7 +17519,7 @@ } }, { - "accuracy": 0.9975975535344332, + "accuracy": 0.9975970382802188, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -17561,7 +17561,7 @@ ], "model.layers.9.self_attn": [ { - "accuracy": 0.9156324192881584, + "accuracy": 0.9154471009969711, "total_bits": 30308928, "q_proj": { "group_size": { @@ -17625,7 +17625,7 @@ } }, { - "accuracy": 0.918098546564579, + "accuracy": 0.9180809780955315, "total_bits": 31455808, "q_proj": { "group_size": { @@ -17689,7 +17689,7 @@ } }, { - "accuracy": 0.9291017279028893, + "accuracy": 0.9292140081524849, "total_bits": 33412832, "q_proj": { "group_size": { @@ -17753,7 +17753,7 @@ } }, { - "accuracy": 0.9474808238446712, + "accuracy": 0.9475394748151302, "total_bits": 37983200, "q_proj": { "group_size": { @@ -17817,7 +17817,7 @@ } }, { - "accuracy": 0.9566572569310665, + "accuracy": 0.9566517025232315, "total_bits": 44838176, "q_proj": { "group_size": { @@ -17881,7 +17881,7 @@ } }, { - "accuracy": 0.9572707749903202, + "accuracy": 0.9572483897209167, "total_bits": 44912768, "q_proj": { "group_size": { @@ -17945,7 +17945,7 @@ } }, { - "accuracy": 0.9765746779739857, + "accuracy": 0.9765565041452646, "total_bits": 57355552, "q_proj": { "group_size": { @@ -17997,7 +17997,7 @@ } }, { - "accuracy": 0.9771516304463148, + "accuracy": 0.9771292954683304, "total_bits": 57430144, "q_proj": { "group_size": { @@ -18049,7 +18049,7 @@ } }, { - "accuracy": 0.9786375071853399, + "accuracy": 0.9786593876779079, "total_bits": 57950464, "q_proj": { "group_size": { @@ -18101,7 +18101,7 @@ } }, { - "accuracy": 0.9796764496713877, + "accuracy": 0.9796735905110836, "total_bits": 58692736, "q_proj": { "group_size": { @@ -18153,7 +18153,7 @@ } }, { - "accuracy": 0.9782940931618214, + "accuracy": 0.9782825838774443, "total_bits": 59068544, "q_proj": { "group_size": { @@ -18217,7 +18217,7 @@ } }, { - "accuracy": 0.9796239528805017, + "accuracy": 0.9796628952026367, "total_bits": 59588864, "q_proj": { "group_size": { @@ -18281,7 +18281,7 @@ } }, { - "accuracy": 0.9821039438247681, + "accuracy": 0.9821524657309055, "total_bits": 61536832, "q_proj": { "group_size": { @@ -18342,7 +18342,7 @@ } }, { - "accuracy": 0.9833755418658257, + "accuracy": 0.9833454601466656, "total_bits": 62354560, "q_proj": { "group_size": { @@ -18403,7 +18403,7 @@ } }, { - "accuracy": 0.9902575109153986, + "accuracy": 0.9902483588084579, "total_bits": 75246880, "q_proj": { "group_size": { @@ -18464,7 +18464,7 @@ } }, { - "accuracy": 0.9916903469711542, + "accuracy": 0.9916817611083388, "total_bits": 76510336, "q_proj": { "group_size": { @@ -18525,7 +18525,7 @@ } }, { - "accuracy": 0.993913528509438, + "accuracy": 0.9939101827330887, "total_bits": 85667104, "q_proj": { "group_size": { @@ -18577,7 +18577,7 @@ } }, { - "accuracy": 0.995829266961664, + "accuracy": 0.99583475664258, "total_bits": 91722880, "q_proj": { "group_size": { @@ -18629,7 +18629,7 @@ } }, { - "accuracy": 0.9984044209122658, + "accuracy": 0.9984046985628083, "total_bits": 113978656, "q_proj": { "group_size": { @@ -18683,7 +18683,7 @@ ], "model.layers.9.mlp": [ { - "accuracy": 0.8979872688651085, + "accuracy": 0.8980022817850113, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -18735,7 +18735,7 @@ } }, { - "accuracy": 0.9010738134384155, + "accuracy": 0.9009037613868713, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -18787,7 +18787,7 @@ } }, { - "accuracy": 0.9138390645384789, + "accuracy": 0.9135458469390869, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -18836,7 +18836,7 @@ } }, { - "accuracy": 0.9180445671081543, + "accuracy": 0.9177292063832283, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -18885,7 +18885,7 @@ } }, { - "accuracy": 0.9479075893759727, + "accuracy": 0.9478593431413174, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -18937,7 +18937,7 @@ } }, { - "accuracy": 0.9523597359657288, + "accuracy": 0.9523407816886902, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -18989,7 +18989,7 @@ } }, { - "accuracy": 0.9581680968403816, + "accuracy": 0.958098616451025, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -19038,7 +19038,7 @@ } }, { - "accuracy": 0.9725048933178186, + "accuracy": 0.9725117534399033, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -19081,7 +19081,7 @@ } }, { - "accuracy": 0.9750282820314169, + "accuracy": 0.975043885409832, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -19124,7 +19124,7 @@ } }, { - "accuracy": 0.9732320122420788, + "accuracy": 0.9732519965618849, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -19176,7 +19176,7 @@ } }, { - "accuracy": 0.9767951015383005, + "accuracy": 0.9767921902239323, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -19228,7 +19228,7 @@ } }, { - "accuracy": 0.9863404175266623, + "accuracy": 0.9863471165299416, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -19280,7 +19280,7 @@ } }, { - "accuracy": 0.9883861737325788, + "accuracy": 0.9883871376514435, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -19332,7 +19332,7 @@ } }, { - "accuracy": 0.992430523969233, + "accuracy": 0.9924324788153172, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -19375,7 +19375,7 @@ } }, { - "accuracy": 0.9929885254241526, + "accuracy": 0.9929874055087566, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -19424,7 +19424,7 @@ } }, { - "accuracy": 0.9942358685657382, + "accuracy": 0.9942356436513364, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -19470,7 +19470,7 @@ } }, { - "accuracy": 0.9979169676080346, + "accuracy": 0.9979170782025903, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -19512,7 +19512,7 @@ ], "model.layers.10.self_attn": [ { - "accuracy": 0.9255667924880981, + "accuracy": 0.9254285618662834, "total_bits": 30308928, "q_proj": { "group_size": { @@ -19576,7 +19576,7 @@ } }, { - "accuracy": 0.9285323470830917, + "accuracy": 0.9285455048084259, "total_bits": 31455808, "q_proj": { "group_size": { @@ -19640,7 +19640,7 @@ } }, { - "accuracy": 0.9377534314990044, + "accuracy": 0.9378277622163296, "total_bits": 33412832, "q_proj": { "group_size": { @@ -19704,7 +19704,7 @@ } }, { - "accuracy": 0.9557530134916306, + "accuracy": 0.9558689743280411, "total_bits": 37983200, "q_proj": { "group_size": { @@ -19768,7 +19768,7 @@ } }, { - "accuracy": 0.9622449390590191, + "accuracy": 0.962271124124527, "total_bits": 44838176, "q_proj": { "group_size": { @@ -19832,7 +19832,7 @@ } }, { - "accuracy": 0.9627878665924072, + "accuracy": 0.9627652131021023, "total_bits": 44912768, "q_proj": { "group_size": { @@ -19896,7 +19896,7 @@ } }, { - "accuracy": 0.9794471673667431, + "accuracy": 0.9794409442692995, "total_bits": 57355552, "q_proj": { "group_size": { @@ -19948,7 +19948,7 @@ } }, { - "accuracy": 0.9799556694924831, + "accuracy": 0.9799547977745533, "total_bits": 57430144, "q_proj": { "group_size": { @@ -20000,7 +20000,7 @@ } }, { - "accuracy": 0.9812588579952717, + "accuracy": 0.9812564067542553, "total_bits": 57950464, "q_proj": { "group_size": { @@ -20052,7 +20052,7 @@ } }, { - "accuracy": 0.982187744230032, + "accuracy": 0.98219870403409, "total_bits": 58692736, "q_proj": { "group_size": { @@ -20104,7 +20104,7 @@ } }, { - "accuracy": 0.9810379017144442, + "accuracy": 0.9810190796852112, "total_bits": 59068544, "q_proj": { "group_size": { @@ -20168,7 +20168,7 @@ } }, { - "accuracy": 0.9822507984936237, + "accuracy": 0.9822470918297768, "total_bits": 59588864, "q_proj": { "group_size": { @@ -20232,7 +20232,7 @@ } }, { - "accuracy": 0.9842813797295094, + "accuracy": 0.9842821806669235, "total_bits": 61536832, "q_proj": { "group_size": { @@ -20293,7 +20293,7 @@ } }, { - "accuracy": 0.9854367896914482, + "accuracy": 0.9854378383606672, "total_bits": 62354560, "q_proj": { "group_size": { @@ -20354,7 +20354,7 @@ } }, { - "accuracy": 0.991425178013742, + "accuracy": 0.9914244255051017, "total_bits": 75246880, "q_proj": { "group_size": { @@ -20415,7 +20415,7 @@ } }, { - "accuracy": 0.9926807903684676, + "accuracy": 0.9926830539479852, "total_bits": 76510336, "q_proj": { "group_size": { @@ -20476,7 +20476,7 @@ } }, { - "accuracy": 0.9946406041271985, + "accuracy": 0.994636666495353, "total_bits": 85667104, "q_proj": { "group_size": { @@ -20528,7 +20528,7 @@ } }, { - "accuracy": 0.9963198550976813, + "accuracy": 0.9963192888535559, "total_bits": 91722880, "q_proj": { "group_size": { @@ -20580,7 +20580,7 @@ } }, { - "accuracy": 0.9985976866446435, + "accuracy": 0.9985979022458196, "total_bits": 113978656, "q_proj": { "group_size": { @@ -20634,7 +20634,7 @@ ], "model.layers.10.mlp": [ { - "accuracy": 0.9072035849094391, + "accuracy": 0.9071538224816322, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -20686,7 +20686,7 @@ } }, { - "accuracy": 0.9100625887513161, + "accuracy": 0.9100388288497925, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -20738,7 +20738,7 @@ } }, { - "accuracy": 0.9221524447202682, + "accuracy": 0.922150731086731, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -20787,7 +20787,7 @@ } }, { - "accuracy": 0.9259821102023125, + "accuracy": 0.9259923249483109, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -20836,7 +20836,7 @@ } }, { - "accuracy": 0.9525629505515099, + "accuracy": 0.9525576233863831, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -20888,7 +20888,7 @@ } }, { - "accuracy": 0.9565679579973221, + "accuracy": 0.9565683789551258, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -20940,7 +20940,7 @@ } }, { - "accuracy": 0.9619633853435516, + "accuracy": 0.9619757160544395, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -20989,7 +20989,7 @@ } }, { - "accuracy": 0.975079670548439, + "accuracy": 0.9750785324722528, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -21032,7 +21032,7 @@ } }, { - "accuracy": 0.9773116856813431, + "accuracy": 0.9773098323494196, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -21075,7 +21075,7 @@ } }, { - "accuracy": 0.9756849519908428, + "accuracy": 0.9756814874708652, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -21127,7 +21127,7 @@ } }, { - "accuracy": 0.9788219351321459, + "accuracy": 0.9788009151816368, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -21179,7 +21179,7 @@ } }, { - "accuracy": 0.9875948084518313, + "accuracy": 0.9875935269519687, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -21231,7 +21231,7 @@ } }, { - "accuracy": 0.9894048757851124, + "accuracy": 0.9894073354080319, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -21283,7 +21283,7 @@ } }, { - "accuracy": 0.993154349271208, + "accuracy": 0.9931564317084849, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -21326,7 +21326,7 @@ } }, { - "accuracy": 0.9936201777309179, + "accuracy": 0.9936129190027714, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -21375,7 +21375,7 @@ } }, { - "accuracy": 0.9947737217880785, + "accuracy": 0.9947650441899896, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -21421,7 +21421,7 @@ } }, { - "accuracy": 0.9981311410665512, + "accuracy": 0.9981314450269565, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -21463,7 +21463,7 @@ ], "model.layers.11.self_attn": [ { - "accuracy": 0.9352168366312981, + "accuracy": 0.9351540505886078, "total_bits": 30308928, "q_proj": { "group_size": { @@ -21527,7 +21527,7 @@ } }, { - "accuracy": 0.9374474063515663, + "accuracy": 0.9374874904751778, "total_bits": 31455808, "q_proj": { "group_size": { @@ -21591,7 +21591,7 @@ } }, { - "accuracy": 0.9444124512374401, + "accuracy": 0.9444890134036541, "total_bits": 33412832, "q_proj": { "group_size": { @@ -21655,7 +21655,7 @@ } }, { - "accuracy": 0.9625572040677071, + "accuracy": 0.9625845514237881, "total_bits": 37983200, "q_proj": { "group_size": { @@ -21719,7 +21719,7 @@ } }, { - "accuracy": 0.9669128507375717, + "accuracy": 0.9669028967618942, "total_bits": 44838176, "q_proj": { "group_size": { @@ -21783,7 +21783,7 @@ } }, { - "accuracy": 0.967328879982233, + "accuracy": 0.9672983437776566, "total_bits": 44912768, "q_proj": { "group_size": { @@ -21847,7 +21847,7 @@ } }, { - "accuracy": 0.9819058496505022, + "accuracy": 0.9818990211933851, "total_bits": 57355552, "q_proj": { "group_size": { @@ -21899,7 +21899,7 @@ } }, { - "accuracy": 0.9823221303522587, + "accuracy": 0.9823174811899662, "total_bits": 57430144, "q_proj": { "group_size": { @@ -21951,7 +21951,7 @@ } }, { - "accuracy": 0.9834885075688362, + "accuracy": 0.9834803491830826, "total_bits": 57950464, "q_proj": { "group_size": { @@ -22003,7 +22003,7 @@ } }, { - "accuracy": 0.9843579828739166, + "accuracy": 0.9843547828495502, "total_bits": 58692736, "q_proj": { "group_size": { @@ -22055,7 +22055,7 @@ } }, { - "accuracy": 0.9833086878061295, + "accuracy": 0.9833086282014847, "total_bits": 59068544, "q_proj": { "group_size": { @@ -22119,7 +22119,7 @@ } }, { - "accuracy": 0.9843683261424303, + "accuracy": 0.9843555800616741, "total_bits": 59588864, "q_proj": { "group_size": { @@ -22183,7 +22183,7 @@ } }, { - "accuracy": 0.9858474927023053, + "accuracy": 0.9858486671000719, "total_bits": 61536832, "q_proj": { "group_size": { @@ -22244,7 +22244,7 @@ } }, { - "accuracy": 0.9868558822199702, + "accuracy": 0.9868420660495758, "total_bits": 62354560, "q_proj": { "group_size": { @@ -22305,7 +22305,7 @@ } }, { - "accuracy": 0.9922798662446439, + "accuracy": 0.9922812515869737, "total_bits": 75246880, "q_proj": { "group_size": { @@ -22366,7 +22366,7 @@ } }, { - "accuracy": 0.9933844502083957, + "accuracy": 0.9933867929503322, "total_bits": 76510336, "q_proj": { "group_size": { @@ -22427,7 +22427,7 @@ } }, { - "accuracy": 0.9952675346285105, + "accuracy": 0.9952673111110926, "total_bits": 85667104, "q_proj": { "group_size": { @@ -22479,7 +22479,7 @@ } }, { - "accuracy": 0.9966305803973228, + "accuracy": 0.9966264350805432, "total_bits": 91722880, "q_proj": { "group_size": { @@ -22531,7 +22531,7 @@ } }, { - "accuracy": 0.9987549100769684, + "accuracy": 0.9987560133449733, "total_bits": 113978656, "q_proj": { "group_size": { @@ -22585,7 +22585,7 @@ ], "model.layers.11.mlp": [ { - "accuracy": 0.922992967069149, + "accuracy": 0.9229258298873901, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -22637,7 +22637,7 @@ } }, { - "accuracy": 0.9253613352775574, + "accuracy": 0.9253562912344933, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -22689,7 +22689,7 @@ } }, { - "accuracy": 0.9349236041307449, + "accuracy": 0.9348956197500229, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -22738,7 +22738,7 @@ } }, { - "accuracy": 0.9380608797073364, + "accuracy": 0.9380255676805973, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -22787,7 +22787,7 @@ } }, { - "accuracy": 0.9607282392680645, + "accuracy": 0.9607758708298206, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -22839,7 +22839,7 @@ } }, { - "accuracy": 0.9641470424830914, + "accuracy": 0.964177817106247, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -22891,7 +22891,7 @@ } }, { - "accuracy": 0.9683838337659836, + "accuracy": 0.9684282056987286, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -22940,7 +22940,7 @@ } }, { - "accuracy": 0.9791794028133154, + "accuracy": 0.9791763704270124, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -22983,7 +22983,7 @@ } }, { - "accuracy": 0.9810750614851713, + "accuracy": 0.9810761008411646, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -23026,7 +23026,7 @@ } }, { - "accuracy": 0.9798438455909491, + "accuracy": 0.9798630569130182, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -23078,7 +23078,7 @@ } }, { - "accuracy": 0.9825253784656525, + "accuracy": 0.9825179129838943, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -23130,7 +23130,7 @@ } }, { - "accuracy": 0.9897021958604455, + "accuracy": 0.9896943140774965, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -23182,7 +23182,7 @@ } }, { - "accuracy": 0.9912520227953792, + "accuracy": 0.9912572083994746, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -23234,7 +23234,7 @@ } }, { - "accuracy": 0.994236150290817, + "accuracy": 0.9942377745173872, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -23277,7 +23277,7 @@ } }, { - "accuracy": 0.9947156864218414, + "accuracy": 0.9947184869088233, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -23326,7 +23326,7 @@ } }, { - "accuracy": 0.9956142762675881, + "accuracy": 0.9956165258772671, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -23372,7 +23372,7 @@ } }, { - "accuracy": 0.9984074216336012, + "accuracy": 0.9984078536508605, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -23414,7 +23414,7 @@ ], "model.layers.12.self_attn": [ { - "accuracy": 0.918254628777504, + "accuracy": 0.9183487147092819, "total_bits": 30308928, "q_proj": { "group_size": { @@ -23478,7 +23478,7 @@ } }, { - "accuracy": 0.9215041920542717, + "accuracy": 0.9214219599962234, "total_bits": 31455808, "q_proj": { "group_size": { @@ -23542,7 +23542,7 @@ } }, { - "accuracy": 0.9291860163211823, + "accuracy": 0.9291634485125542, "total_bits": 33412832, "q_proj": { "group_size": { @@ -23606,7 +23606,7 @@ } }, { - "accuracy": 0.9523625448346138, + "accuracy": 0.9523275718092918, "total_bits": 37983200, "q_proj": { "group_size": { @@ -23670,7 +23670,7 @@ } }, { - "accuracy": 0.9582577794790268, + "accuracy": 0.958238810300827, "total_bits": 44838176, "q_proj": { "group_size": { @@ -23734,7 +23734,7 @@ } }, { - "accuracy": 0.9587131217122078, + "accuracy": 0.9586759731173515, "total_bits": 44912768, "q_proj": { "group_size": { @@ -23798,7 +23798,7 @@ } }, { - "accuracy": 0.9767006933689117, + "accuracy": 0.9766934681683779, "total_bits": 57355552, "q_proj": { "group_size": { @@ -23850,7 +23850,7 @@ } }, { - "accuracy": 0.9771417211741209, + "accuracy": 0.9771546181291342, "total_bits": 57430144, "q_proj": { "group_size": { @@ -23902,7 +23902,7 @@ } }, { - "accuracy": 0.9786731544882059, + "accuracy": 0.978665629401803, "total_bits": 57950464, "q_proj": { "group_size": { @@ -23954,7 +23954,7 @@ } }, { - "accuracy": 0.9797742627561092, + "accuracy": 0.9797797929495573, "total_bits": 58692736, "q_proj": { "group_size": { @@ -24006,7 +24006,7 @@ } }, { - "accuracy": 0.9788827616721392, + "accuracy": 0.9788805264979601, "total_bits": 59068544, "q_proj": { "group_size": { @@ -24070,7 +24070,7 @@ } }, { - "accuracy": 0.980213237926364, + "accuracy": 0.9802060816437006, "total_bits": 59588864, "q_proj": { "group_size": { @@ -24134,7 +24134,7 @@ } }, { - "accuracy": 0.9817759990692139, + "accuracy": 0.9817687887698412, "total_bits": 61536832, "q_proj": { "group_size": { @@ -24195,7 +24195,7 @@ } }, { - "accuracy": 0.9829946663230658, + "accuracy": 0.9829999972134829, "total_bits": 62354560, "q_proj": { "group_size": { @@ -24256,7 +24256,7 @@ } }, { - "accuracy": 0.9900454664602876, + "accuracy": 0.9900399595499039, "total_bits": 75246880, "q_proj": { "group_size": { @@ -24317,7 +24317,7 @@ } }, { - "accuracy": 0.9914399096742272, + "accuracy": 0.9914370570331812, "total_bits": 76510336, "q_proj": { "group_size": { @@ -24378,7 +24378,7 @@ } }, { - "accuracy": 0.9938779231160879, + "accuracy": 0.9938817266374826, "total_bits": 85667104, "q_proj": { "group_size": { @@ -24430,7 +24430,7 @@ } }, { - "accuracy": 0.9955275356769562, + "accuracy": 0.9955261144787073, "total_bits": 91722880, "q_proj": { "group_size": { @@ -24482,7 +24482,7 @@ } }, { - "accuracy": 0.998387377592735, + "accuracy": 0.998388102161698, "total_bits": 113978656, "q_proj": { "group_size": { @@ -24536,7 +24536,7 @@ ], "model.layers.12.mlp": [ { - "accuracy": 0.9067049399018288, + "accuracy": 0.9067919254302979, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -24588,7 +24588,7 @@ } }, { - "accuracy": 0.9094245880842209, + "accuracy": 0.9094521254301071, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -24640,7 +24640,7 @@ } }, { - "accuracy": 0.9205821231007576, + "accuracy": 0.9205865487456322, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -24689,7 +24689,7 @@ } }, { - "accuracy": 0.924173966050148, + "accuracy": 0.9241846203804016, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -24738,7 +24738,7 @@ } }, { - "accuracy": 0.9520841389894485, + "accuracy": 0.9520521014928818, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -24790,7 +24790,7 @@ } }, { - "accuracy": 0.9562458842992783, + "accuracy": 0.9562351442873478, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -24842,7 +24842,7 @@ } }, { - "accuracy": 0.9611620157957077, + "accuracy": 0.961151484400034, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -24891,7 +24891,7 @@ } }, { - "accuracy": 0.9748448915779591, + "accuracy": 0.9748354908078909, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -24934,7 +24934,7 @@ } }, { - "accuracy": 0.9771162606775761, + "accuracy": 0.9771051350980997, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -24977,7 +24977,7 @@ } }, { - "accuracy": 0.975365836173296, + "accuracy": 0.9753743782639503, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -25029,7 +25029,7 @@ } }, { - "accuracy": 0.9786117561161518, + "accuracy": 0.9786114767193794, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -25081,7 +25081,7 @@ } }, { - "accuracy": 0.9873783187940717, + "accuracy": 0.9873899882659316, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -25133,7 +25133,7 @@ } }, { - "accuracy": 0.9893058259040117, + "accuracy": 0.9893016312271357, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -25185,7 +25185,7 @@ } }, { - "accuracy": 0.9930427521467209, + "accuracy": 0.993039789609611, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -25228,7 +25228,7 @@ } }, { - "accuracy": 0.9935374888591468, + "accuracy": 0.9935355954803526, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -25277,7 +25277,7 @@ } }, { - "accuracy": 0.9945738040842116, + "accuracy": 0.9945709388703108, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -25323,7 +25323,7 @@ } }, { - "accuracy": 0.9981157196452841, + "accuracy": 0.9981151465326548, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -25365,7 +25365,7 @@ ], "model.layers.13.self_attn": [ { - "accuracy": 0.914382241666317, + "accuracy": 0.9144919887185097, "total_bits": 30308928, "q_proj": { "group_size": { @@ -25429,7 +25429,7 @@ } }, { - "accuracy": 0.9174628779292107, + "accuracy": 0.9174542501568794, "total_bits": 31455808, "q_proj": { "group_size": { @@ -25493,7 +25493,7 @@ } }, { - "accuracy": 0.9295148551464081, + "accuracy": 0.9295745342969894, "total_bits": 33412832, "q_proj": { "group_size": { @@ -25557,7 +25557,7 @@ } }, { - "accuracy": 0.9491968564689159, + "accuracy": 0.9492183402180672, "total_bits": 37983200, "q_proj": { "group_size": { @@ -25621,7 +25621,7 @@ } }, { - "accuracy": 0.955919798463583, + "accuracy": 0.9559485502541065, "total_bits": 44838176, "q_proj": { "group_size": { @@ -25685,7 +25685,7 @@ } }, { - "accuracy": 0.9567287154495716, + "accuracy": 0.9566921889781952, "total_bits": 44912768, "q_proj": { "group_size": { @@ -25749,7 +25749,7 @@ } }, { - "accuracy": 0.9756081849336624, + "accuracy": 0.9756113328039646, "total_bits": 57355552, "q_proj": { "group_size": { @@ -25801,7 +25801,7 @@ } }, { - "accuracy": 0.9762581083923578, + "accuracy": 0.9762618523091078, "total_bits": 57430144, "q_proj": { "group_size": { @@ -25853,7 +25853,7 @@ } }, { - "accuracy": 0.9783238749951124, + "accuracy": 0.9782921988517046, "total_bits": 57950464, "q_proj": { "group_size": { @@ -25905,7 +25905,7 @@ } }, { - "accuracy": 0.9795089289546013, + "accuracy": 0.9794944487512112, "total_bits": 58692736, "q_proj": { "group_size": { @@ -25957,7 +25957,7 @@ } }, { - "accuracy": 0.9779392145574093, + "accuracy": 0.9778795074671507, "total_bits": 59068544, "q_proj": { "group_size": { @@ -26021,7 +26021,7 @@ } }, { - "accuracy": 0.9795319139957428, + "accuracy": 0.979520695284009, "total_bits": 59588864, "q_proj": { "group_size": { @@ -26085,7 +26085,7 @@ } }, { - "accuracy": 0.9819448851048946, + "accuracy": 0.9819323495030403, "total_bits": 61536832, "q_proj": { "group_size": { @@ -26146,7 +26146,7 @@ } }, { - "accuracy": 0.9835209306329489, + "accuracy": 0.9835347458720207, "total_bits": 62354560, "q_proj": { "group_size": { @@ -26207,7 +26207,7 @@ } }, { - "accuracy": 0.9901209101080894, + "accuracy": 0.9901253506541252, "total_bits": 75246880, "q_proj": { "group_size": { @@ -26268,7 +26268,7 @@ } }, { - "accuracy": 0.9917429555207491, + "accuracy": 0.9917331263422966, "total_bits": 76510336, "q_proj": { "group_size": { @@ -26329,7 +26329,7 @@ } }, { - "accuracy": 0.9937325660139322, + "accuracy": 0.9937320556491613, "total_bits": 85667104, "q_proj": { "group_size": { @@ -26381,7 +26381,7 @@ } }, { - "accuracy": 0.9958382518962026, + "accuracy": 0.9958483558148146, "total_bits": 91722880, "q_proj": { "group_size": { @@ -26433,7 +26433,7 @@ } }, { - "accuracy": 0.9983558729290962, + "accuracy": 0.9983553681522608, "total_bits": 113978656, "q_proj": { "group_size": { @@ -26487,7 +26487,7 @@ ], "model.layers.13.mlp": [ { - "accuracy": 0.9058817103505135, + "accuracy": 0.9057564288377762, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -26539,7 +26539,7 @@ } }, { - "accuracy": 0.9087112993001938, + "accuracy": 0.9086236134171486, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -26591,7 +26591,7 @@ } }, { - "accuracy": 0.9195607155561447, + "accuracy": 0.9196574911475182, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -26640,7 +26640,7 @@ } }, { - "accuracy": 0.923167884349823, + "accuracy": 0.9232490658760071, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -26689,7 +26689,7 @@ } }, { - "accuracy": 0.9515799544751644, + "accuracy": 0.9516337104141712, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -26741,7 +26741,7 @@ } }, { - "accuracy": 0.9560265839099884, + "accuracy": 0.9560323432087898, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -26793,7 +26793,7 @@ } }, { - "accuracy": 0.9607695490121841, + "accuracy": 0.9607681259512901, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -26842,7 +26842,7 @@ } }, { - "accuracy": 0.9745559766888618, + "accuracy": 0.9745499193668365, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -26885,7 +26885,7 @@ } }, { - "accuracy": 0.976823752745986, + "accuracy": 0.9768188558518887, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -26928,7 +26928,7 @@ } }, { - "accuracy": 0.9750555753707886, + "accuracy": 0.9750737082213163, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -26980,7 +26980,7 @@ } }, { - "accuracy": 0.978454452008009, + "accuracy": 0.9784559179097414, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -27032,7 +27032,7 @@ } }, { - "accuracy": 0.9872105978429317, + "accuracy": 0.9872140353545547, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -27084,7 +27084,7 @@ } }, { - "accuracy": 0.9891844298690557, + "accuracy": 0.9891853602603078, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -27136,7 +27136,7 @@ } }, { - "accuracy": 0.992884736508131, + "accuracy": 0.9928796184249222, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -27179,7 +27179,7 @@ } }, { - "accuracy": 0.9934014892205596, + "accuracy": 0.9934069863520563, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -27228,7 +27228,7 @@ } }, { - "accuracy": 0.9944165083579719, + "accuracy": 0.9944231859408319, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -27274,7 +27274,7 @@ } }, { - "accuracy": 0.9980151185300201, + "accuracy": 0.9980159213300794, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -27316,7 +27316,7 @@ ], "model.layers.14.self_attn": [ { - "accuracy": 0.9055267497897148, + "accuracy": 0.9052515029907227, "total_bits": 30308928, "q_proj": { "group_size": { @@ -27380,7 +27380,7 @@ } }, { - "accuracy": 0.9082472920417786, + "accuracy": 0.9081130772829056, "total_bits": 31455808, "q_proj": { "group_size": { @@ -27444,7 +27444,7 @@ } }, { - "accuracy": 0.9182290732860565, + "accuracy": 0.9183298647403717, "total_bits": 33412832, "q_proj": { "group_size": { @@ -27508,7 +27508,7 @@ } }, { - "accuracy": 0.9408689215779305, + "accuracy": 0.9409149326384068, "total_bits": 37983200, "q_proj": { "group_size": { @@ -27572,7 +27572,7 @@ } }, { - "accuracy": 0.9510448761284351, + "accuracy": 0.9511469714343548, "total_bits": 44838176, "q_proj": { "group_size": { @@ -27636,7 +27636,7 @@ } }, { - "accuracy": 0.9517546966671944, + "accuracy": 0.9518323987722397, "total_bits": 44912768, "q_proj": { "group_size": { @@ -27700,7 +27700,7 @@ } }, { - "accuracy": 0.9723473638296127, + "accuracy": 0.9723943509161472, "total_bits": 57355552, "q_proj": { "group_size": { @@ -27752,7 +27752,7 @@ } }, { - "accuracy": 0.9730114750564098, + "accuracy": 0.9731061886996031, "total_bits": 57430144, "q_proj": { "group_size": { @@ -27804,7 +27804,7 @@ } }, { - "accuracy": 0.9746960289776325, + "accuracy": 0.9747657887637615, "total_bits": 57950464, "q_proj": { "group_size": { @@ -27856,7 +27856,7 @@ } }, { - "accuracy": 0.9767959043383598, + "accuracy": 0.9768108148127794, "total_bits": 58692736, "q_proj": { "group_size": { @@ -27908,7 +27908,7 @@ } }, { - "accuracy": 0.9753426536917686, + "accuracy": 0.975377295166254, "total_bits": 59068544, "q_proj": { "group_size": { @@ -27972,7 +27972,7 @@ } }, { - "accuracy": 0.9770860001444817, + "accuracy": 0.9770306386053562, "total_bits": 59588864, "q_proj": { "group_size": { @@ -28036,7 +28036,7 @@ } }, { - "accuracy": 0.979099478572607, + "accuracy": 0.9790472611784935, "total_bits": 61536832, "q_proj": { "group_size": { @@ -28097,7 +28097,7 @@ } }, { - "accuracy": 0.9806705601513386, + "accuracy": 0.9806842133402824, "total_bits": 62354560, "q_proj": { "group_size": { @@ -28158,7 +28158,7 @@ } }, { - "accuracy": 0.9885000661015511, + "accuracy": 0.9885049602016807, "total_bits": 75246880, "q_proj": { "group_size": { @@ -28219,7 +28219,7 @@ } }, { - "accuracy": 0.9902753746137023, + "accuracy": 0.9902975112199783, "total_bits": 76510336, "q_proj": { "group_size": { @@ -28280,7 +28280,7 @@ } }, { - "accuracy": 0.992810903582722, + "accuracy": 0.9927993132732809, "total_bits": 85667104, "q_proj": { "group_size": { @@ -28332,7 +28332,7 @@ } }, { - "accuracy": 0.9949225764721632, + "accuracy": 0.994893008377403, "total_bits": 91722880, "q_proj": { "group_size": { @@ -28384,7 +28384,7 @@ } }, { - "accuracy": 0.9981167989317328, + "accuracy": 0.998116486473009, "total_bits": 113978656, "q_proj": { "group_size": { @@ -28438,7 +28438,7 @@ ], "model.layers.14.mlp": [ { - "accuracy": 0.9059914350509644, + "accuracy": 0.90602907538414, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -28490,7 +28490,7 @@ } }, { - "accuracy": 0.9088080897927284, + "accuracy": 0.9087480306625366, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -28542,7 +28542,7 @@ } }, { - "accuracy": 0.9203511402010918, + "accuracy": 0.9203285947442055, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -28591,7 +28591,7 @@ } }, { - "accuracy": 0.9241925030946732, + "accuracy": 0.9241577237844467, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -28640,7 +28640,7 @@ } }, { - "accuracy": 0.9516731016337872, + "accuracy": 0.951668631285429, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -28692,7 +28692,7 @@ } }, { - "accuracy": 0.9560071490705013, + "accuracy": 0.9559860564768314, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -28744,7 +28744,7 @@ } }, { - "accuracy": 0.9610604681074619, + "accuracy": 0.9610610008239746, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -28793,7 +28793,7 @@ } }, { - "accuracy": 0.9745946303009987, + "accuracy": 0.9746211394667625, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -28836,7 +28836,7 @@ } }, { - "accuracy": 0.9768316633999348, + "accuracy": 0.9768539238721132, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -28879,7 +28879,7 @@ } }, { - "accuracy": 0.9751278180629015, + "accuracy": 0.9751336295157671, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -28931,7 +28931,7 @@ } }, { - "accuracy": 0.9784498494118452, + "accuracy": 0.978433957323432, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -28983,7 +28983,7 @@ } }, { - "accuracy": 0.9872479913756251, + "accuracy": 0.9872386232018471, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -29035,7 +29035,7 @@ } }, { - "accuracy": 0.9891970083117485, + "accuracy": 0.98919727653265, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -29087,7 +29087,7 @@ } }, { - "accuracy": 0.992910641245544, + "accuracy": 0.9929077257402241, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -29130,7 +29130,7 @@ } }, { - "accuracy": 0.9934103316627443, + "accuracy": 0.99340810906142, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -29179,7 +29179,7 @@ } }, { - "accuracy": 0.9944956391118467, + "accuracy": 0.9944933266378939, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -29225,7 +29225,7 @@ } }, { - "accuracy": 0.9980454861652106, + "accuracy": 0.9980455068871379, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -29267,7 +29267,7 @@ ], "model.layers.15.self_attn": [ { - "accuracy": 0.8662237524986267, + "accuracy": 0.8660960495471954, "total_bits": 30308928, "q_proj": { "group_size": { @@ -29331,7 +29331,7 @@ } }, { - "accuracy": 0.8720844835042953, + "accuracy": 0.8721434473991394, "total_bits": 31455808, "q_proj": { "group_size": { @@ -29395,7 +29395,7 @@ } }, { - "accuracy": 0.8906018137931824, + "accuracy": 0.8906832113862038, "total_bits": 33412832, "q_proj": { "group_size": { @@ -29459,7 +29459,7 @@ } }, { - "accuracy": 0.922219954431057, + "accuracy": 0.9223601892590523, "total_bits": 37983200, "q_proj": { "group_size": { @@ -29523,7 +29523,7 @@ } }, { - "accuracy": 0.9315391108393669, + "accuracy": 0.9315524771809578, "total_bits": 44838176, "q_proj": { "group_size": { @@ -29587,7 +29587,7 @@ } }, { - "accuracy": 0.9328188821673393, + "accuracy": 0.9328577220439911, "total_bits": 44912768, "q_proj": { "group_size": { @@ -29651,7 +29651,7 @@ } }, { - "accuracy": 0.9618901945650578, + "accuracy": 0.9618603363633156, "total_bits": 57355552, "q_proj": { "group_size": { @@ -29703,7 +29703,7 @@ } }, { - "accuracy": 0.9632286131381989, + "accuracy": 0.9631414413452148, "total_bits": 57430144, "q_proj": { "group_size": { @@ -29755,7 +29755,7 @@ } }, { - "accuracy": 0.9660533033311367, + "accuracy": 0.9660372957587242, "total_bits": 57950464, "q_proj": { "group_size": { @@ -29807,7 +29807,7 @@ } }, { - "accuracy": 0.9675209484994411, + "accuracy": 0.9675589837133884, "total_bits": 58692736, "q_proj": { "group_size": { @@ -29859,7 +29859,7 @@ } }, { - "accuracy": 0.9656884074211121, + "accuracy": 0.9657478630542755, "total_bits": 59068544, "q_proj": { "group_size": { @@ -29923,7 +29923,7 @@ } }, { - "accuracy": 0.9681769870221615, + "accuracy": 0.9681409150362015, "total_bits": 59588864, "q_proj": { "group_size": { @@ -29987,7 +29987,7 @@ } }, { - "accuracy": 0.9722221530973911, + "accuracy": 0.9722285084426403, "total_bits": 61536832, "q_proj": { "group_size": { @@ -30048,7 +30048,7 @@ } }, { - "accuracy": 0.9743126630783081, + "accuracy": 0.9743023552000523, "total_bits": 62354560, "q_proj": { "group_size": { @@ -30109,7 +30109,7 @@ } }, { - "accuracy": 0.9847090318799019, + "accuracy": 0.984696488827467, "total_bits": 75246880, "q_proj": { "group_size": { @@ -30170,7 +30170,7 @@ } }, { - "accuracy": 0.987136579118669, + "accuracy": 0.9871381670236588, "total_bits": 76510336, "q_proj": { "group_size": { @@ -30231,7 +30231,7 @@ } }, { - "accuracy": 0.9900257075205445, + "accuracy": 0.9900241773575544, "total_bits": 85667104, "q_proj": { "group_size": { @@ -30283,7 +30283,7 @@ } }, { - "accuracy": 0.993544745258987, + "accuracy": 0.9935367675498128, "total_bits": 91722880, "q_proj": { "group_size": { @@ -30335,7 +30335,7 @@ } }, { - "accuracy": 0.9974129700567573, + "accuracy": 0.9974108194001019, "total_bits": 113978656, "q_proj": { "group_size": { @@ -30389,7 +30389,7 @@ ], "model.layers.15.mlp": [ { - "accuracy": 0.888176292181015, + "accuracy": 0.8881220296025276, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -30441,7 +30441,7 @@ } }, { - "accuracy": 0.8917654976248741, + "accuracy": 0.8917921856045723, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -30493,7 +30493,7 @@ } }, { - "accuracy": 0.9060942456126213, + "accuracy": 0.9061857014894485, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -30542,7 +30542,7 @@ } }, { - "accuracy": 0.9107229635119438, + "accuracy": 0.9108146727085114, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -30591,7 +30591,7 @@ } }, { - "accuracy": 0.9432618841528893, + "accuracy": 0.9432695247232914, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -30643,7 +30643,7 @@ } }, { - "accuracy": 0.9481836818158627, + "accuracy": 0.9481980800628662, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -30695,7 +30695,7 @@ } }, { - "accuracy": 0.9543357789516449, + "accuracy": 0.9543447978794575, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -30744,7 +30744,7 @@ } }, { - "accuracy": 0.9701154325157404, + "accuracy": 0.9701192807406187, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -30787,7 +30787,7 @@ } }, { - "accuracy": 0.9727210849523544, + "accuracy": 0.9727198947221041, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -30830,7 +30830,7 @@ } }, { - "accuracy": 0.9709262512624264, + "accuracy": 0.9709332976490259, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -30882,7 +30882,7 @@ } }, { - "accuracy": 0.9747244548052549, + "accuracy": 0.9747237414121628, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -30934,7 +30934,7 @@ } }, { - "accuracy": 0.9851207751780748, + "accuracy": 0.9851209698244929, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -30986,7 +30986,7 @@ } }, { - "accuracy": 0.9873405825346708, + "accuracy": 0.9873364437371492, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -31038,7 +31038,7 @@ } }, { - "accuracy": 0.991690007969737, + "accuracy": 0.991686038672924, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -31081,7 +31081,7 @@ } }, { - "accuracy": 0.9923195065930486, + "accuracy": 0.9923256579786539, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -31130,7 +31130,7 @@ } }, { - "accuracy": 0.9936661920510232, + "accuracy": 0.993671263102442, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -31176,7 +31176,7 @@ } }, { - "accuracy": 0.9976954362355173, + "accuracy": 0.9976960332132876, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -31218,7 +31218,7 @@ ], "model.layers.16.self_attn": [ { - "accuracy": 0.8969288393855095, + "accuracy": 0.8970862925052643, "total_bits": 30308928, "q_proj": { "group_size": { @@ -31282,7 +31282,7 @@ } }, { - "accuracy": 0.9022532030940056, + "accuracy": 0.9023451805114746, "total_bits": 31455808, "q_proj": { "group_size": { @@ -31346,7 +31346,7 @@ } }, { - "accuracy": 0.9167034402489662, + "accuracy": 0.9166429415345192, "total_bits": 33412832, "q_proj": { "group_size": { @@ -31410,7 +31410,7 @@ } }, { - "accuracy": 0.9358327612280846, + "accuracy": 0.9357787296175957, "total_bits": 37983200, "q_proj": { "group_size": { @@ -31474,7 +31474,7 @@ } }, { - "accuracy": 0.9472634494304657, + "accuracy": 0.9472717791795731, "total_bits": 44838176, "q_proj": { "group_size": { @@ -31538,7 +31538,7 @@ } }, { - "accuracy": 0.9484786465764046, + "accuracy": 0.9484795853495598, "total_bits": 44912768, "q_proj": { "group_size": { @@ -31602,7 +31602,7 @@ } }, { - "accuracy": 0.9694637320935726, + "accuracy": 0.9694760199636221, "total_bits": 57355552, "q_proj": { "group_size": { @@ -31654,7 +31654,7 @@ } }, { - "accuracy": 0.9708479661494493, + "accuracy": 0.9708113986998796, "total_bits": 57430144, "q_proj": { "group_size": { @@ -31706,7 +31706,7 @@ } }, { - "accuracy": 0.9735263660550117, + "accuracy": 0.9735017456114292, "total_bits": 57950464, "q_proj": { "group_size": { @@ -31758,7 +31758,7 @@ } }, { - "accuracy": 0.9748429581522942, + "accuracy": 0.9747295752167702, "total_bits": 58692736, "q_proj": { "group_size": { @@ -31810,7 +31810,7 @@ } }, { - "accuracy": 0.9736697655171156, + "accuracy": 0.9736849442124367, "total_bits": 59068544, "q_proj": { "group_size": { @@ -31874,7 +31874,7 @@ } }, { - "accuracy": 0.9757199659943581, + "accuracy": 0.9757067002356052, "total_bits": 59588864, "q_proj": { "group_size": { @@ -31938,7 +31938,7 @@ } }, { - "accuracy": 0.9787998888641596, + "accuracy": 0.9787530265748501, "total_bits": 61536832, "q_proj": { "group_size": { @@ -31999,7 +31999,7 @@ } }, { - "accuracy": 0.9805225525051355, + "accuracy": 0.9805027451366186, "total_bits": 62354560, "q_proj": { "group_size": { @@ -32060,7 +32060,7 @@ } }, { - "accuracy": 0.9882316552102566, + "accuracy": 0.9882538206875324, "total_bits": 75246880, "q_proj": { "group_size": { @@ -32121,7 +32121,7 @@ } }, { - "accuracy": 0.9901310745626688, + "accuracy": 0.9901269851252437, "total_bits": 76510336, "q_proj": { "group_size": { @@ -32182,7 +32182,7 @@ } }, { - "accuracy": 0.9920943574979901, + "accuracy": 0.9920847797766328, "total_bits": 85667104, "q_proj": { "group_size": { @@ -32234,7 +32234,7 @@ } }, { - "accuracy": 0.9948356342501938, + "accuracy": 0.9948330130428076, "total_bits": 91722880, "q_proj": { "group_size": { @@ -32286,7 +32286,7 @@ } }, { - "accuracy": 0.9979338457342237, + "accuracy": 0.9979362795129418, "total_bits": 113978656, "q_proj": { "group_size": { @@ -32340,7 +32340,7 @@ ], "model.layers.16.mlp": [ { - "accuracy": 0.8954387977719307, + "accuracy": 0.8952863588929176, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -32392,7 +32392,7 @@ } }, { - "accuracy": 0.8993488028645515, + "accuracy": 0.8992033004760742, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -32444,7 +32444,7 @@ } }, { - "accuracy": 0.9125201851129532, + "accuracy": 0.9123476222157478, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -32493,7 +32493,7 @@ } }, { - "accuracy": 0.916957214474678, + "accuracy": 0.916797399520874, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -32542,7 +32542,7 @@ } }, { - "accuracy": 0.9469495937228203, + "accuracy": 0.9469642452895641, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -32594,7 +32594,7 @@ } }, { - "accuracy": 0.9515363164246082, + "accuracy": 0.951481893658638, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -32646,7 +32646,7 @@ } }, { - "accuracy": 0.9573016315698624, + "accuracy": 0.9572397917509079, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -32695,7 +32695,7 @@ } }, { - "accuracy": 0.97197007201612, + "accuracy": 0.9719776529818773, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -32738,7 +32738,7 @@ } }, { - "accuracy": 0.9743286669254303, + "accuracy": 0.974342368543148, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -32781,7 +32781,7 @@ } }, { - "accuracy": 0.9728555325418711, + "accuracy": 0.9728449210524559, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -32833,7 +32833,7 @@ } }, { - "accuracy": 0.9763589985668659, + "accuracy": 0.9763544015586376, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -32885,7 +32885,7 @@ } }, { - "accuracy": 0.9861350525170565, + "accuracy": 0.9861237006261945, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -32937,7 +32937,7 @@ } }, { - "accuracy": 0.9881947506219149, + "accuracy": 0.9881899319589138, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -32989,7 +32989,7 @@ } }, { - "accuracy": 0.9922279161401093, + "accuracy": 0.9922269973903894, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -33032,7 +33032,7 @@ } }, { - "accuracy": 0.9928275826387107, + "accuracy": 0.9928280608728528, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -33081,7 +33081,7 @@ } }, { - "accuracy": 0.9940253831446171, + "accuracy": 0.9940254967659712, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -33127,7 +33127,7 @@ } }, { - "accuracy": 0.9978487037587911, + "accuracy": 0.9978484844323248, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -33169,7 +33169,7 @@ ], "model.layers.17.self_attn": [ { - "accuracy": 0.8973334357142448, + "accuracy": 0.897542305290699, "total_bits": 30308928, "q_proj": { "group_size": { @@ -33233,7 +33233,7 @@ } }, { - "accuracy": 0.9017208069562912, + "accuracy": 0.9018189832568169, "total_bits": 31455808, "q_proj": { "group_size": { @@ -33297,7 +33297,7 @@ } }, { - "accuracy": 0.9132213443517685, + "accuracy": 0.9131938815116882, "total_bits": 33412832, "q_proj": { "group_size": { @@ -33361,7 +33361,7 @@ } }, { - "accuracy": 0.938893273472786, + "accuracy": 0.9388857409358025, "total_bits": 37983200, "q_proj": { "group_size": { @@ -33425,7 +33425,7 @@ } }, { - "accuracy": 0.9475571848452091, + "accuracy": 0.9475753307342529, "total_bits": 44838176, "q_proj": { "group_size": { @@ -33489,7 +33489,7 @@ } }, { - "accuracy": 0.9484684765338898, + "accuracy": 0.9485311545431614, "total_bits": 44912768, "q_proj": { "group_size": { @@ -33553,7 +33553,7 @@ } }, { - "accuracy": 0.9707033317536116, + "accuracy": 0.970684751868248, "total_bits": 57355552, "q_proj": { "group_size": { @@ -33605,7 +33605,7 @@ } }, { - "accuracy": 0.9716454949229956, + "accuracy": 0.9716228414326906, "total_bits": 57430144, "q_proj": { "group_size": { @@ -33657,7 +33657,7 @@ } }, { - "accuracy": 0.9738726690411568, + "accuracy": 0.973836112767458, "total_bits": 57950464, "q_proj": { "group_size": { @@ -33709,7 +33709,7 @@ } }, { - "accuracy": 0.9751554299145937, + "accuracy": 0.9751142673194408, "total_bits": 58692736, "q_proj": { "group_size": { @@ -33761,7 +33761,7 @@ } }, { - "accuracy": 0.973699688911438, + "accuracy": 0.9736771062016487, "total_bits": 59068544, "q_proj": { "group_size": { @@ -33825,7 +33825,7 @@ } }, { - "accuracy": 0.9755098223686218, + "accuracy": 0.9755203202366829, "total_bits": 59588864, "q_proj": { "group_size": { @@ -33889,7 +33889,7 @@ } }, { - "accuracy": 0.9779850598424673, + "accuracy": 0.9779762793332338, "total_bits": 61536832, "q_proj": { "group_size": { @@ -33950,7 +33950,7 @@ } }, { - "accuracy": 0.9795938245952129, + "accuracy": 0.9795962031930685, "total_bits": 62354560, "q_proj": { "group_size": { @@ -34011,7 +34011,7 @@ } }, { - "accuracy": 0.9879280971363187, + "accuracy": 0.987911774776876, "total_bits": 75246880, "q_proj": { "group_size": { @@ -34072,7 +34072,7 @@ } }, { - "accuracy": 0.989773684181273, + "accuracy": 0.9897674126550555, "total_bits": 76510336, "q_proj": { "group_size": { @@ -34133,7 +34133,7 @@ } }, { - "accuracy": 0.9923741007223725, + "accuracy": 0.9923535925336182, "total_bits": 85667104, "q_proj": { "group_size": { @@ -34185,7 +34185,7 @@ } }, { - "accuracy": 0.9947892669588327, + "accuracy": 0.9947894560173154, "total_bits": 91722880, "q_proj": { "group_size": { @@ -34237,7 +34237,7 @@ } }, { - "accuracy": 0.99802085920237, + "accuracy": 0.998017291072756, "total_bits": 113978656, "q_proj": { "group_size": { @@ -34291,7 +34291,7 @@ ], "model.layers.17.mlp": [ { - "accuracy": 0.8820443674921989, + "accuracy": 0.8820276632905006, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -34343,7 +34343,7 @@ } }, { - "accuracy": 0.8866309821605682, + "accuracy": 0.8865934386849403, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -34395,7 +34395,7 @@ } }, { - "accuracy": 0.9031416177749634, + "accuracy": 0.9032439365983009, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -34444,7 +34444,7 @@ } }, { - "accuracy": 0.9086097851395607, + "accuracy": 0.9086964055895805, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -34493,7 +34493,7 @@ } }, { - "accuracy": 0.9407571367919445, + "accuracy": 0.9407499842345715, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -34545,7 +34545,7 @@ } }, { - "accuracy": 0.9456419833004475, + "accuracy": 0.9456696957349777, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -34597,7 +34597,7 @@ } }, { - "accuracy": 0.9527646824717522, + "accuracy": 0.952761884778738, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -34646,7 +34646,7 @@ } }, { - "accuracy": 0.9684986434876919, + "accuracy": 0.9684955701231956, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -34689,7 +34689,7 @@ } }, { - "accuracy": 0.9712209962308407, + "accuracy": 0.9712157864123583, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -34732,7 +34732,7 @@ } }, { - "accuracy": 0.969695445150137, + "accuracy": 0.969698378816247, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -34784,7 +34784,7 @@ } }, { - "accuracy": 0.973537739366293, + "accuracy": 0.9735384881496429, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -34836,7 +34836,7 @@ } }, { - "accuracy": 0.984518950805068, + "accuracy": 0.9845216348767281, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -34888,7 +34888,7 @@ } }, { - "accuracy": 0.9867853112518787, + "accuracy": 0.9867835436016321, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -34940,7 +34940,7 @@ } }, { - "accuracy": 0.991300848312676, + "accuracy": 0.9913051715120673, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -34983,7 +34983,7 @@ } }, { - "accuracy": 0.9920124607160687, + "accuracy": 0.9920066902413964, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -35032,7 +35032,7 @@ } }, { - "accuracy": 0.9935338972136378, + "accuracy": 0.9935281220823526, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -35078,7 +35078,7 @@ } }, { - "accuracy": 0.9976069342810661, + "accuracy": 0.9976056986488402, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -35120,7 +35120,7 @@ ], "model.layers.18.self_attn": [ { - "accuracy": 0.8961049765348434, + "accuracy": 0.895967960357666, "total_bits": 30308928, "q_proj": { "group_size": { @@ -35184,7 +35184,7 @@ } }, { - "accuracy": 0.9031495451927185, + "accuracy": 0.9031614512205124, "total_bits": 31455808, "q_proj": { "group_size": { @@ -35248,7 +35248,7 @@ } }, { - "accuracy": 0.915522888302803, + "accuracy": 0.9155553206801414, "total_bits": 33412832, "q_proj": { "group_size": { @@ -35312,7 +35312,7 @@ } }, { - "accuracy": 0.9396003894507885, + "accuracy": 0.9396803267300129, "total_bits": 37983200, "q_proj": { "group_size": { @@ -35376,7 +35376,7 @@ } }, { - "accuracy": 0.9481151774525642, + "accuracy": 0.9481409452855587, "total_bits": 44838176, "q_proj": { "group_size": { @@ -35440,7 +35440,7 @@ } }, { - "accuracy": 0.9490873701870441, + "accuracy": 0.9491141587495804, "total_bits": 44912768, "q_proj": { "group_size": { @@ -35504,7 +35504,7 @@ } }, { - "accuracy": 0.9702784214168787, + "accuracy": 0.9702505953609943, "total_bits": 57355552, "q_proj": { "group_size": { @@ -35556,7 +35556,7 @@ } }, { - "accuracy": 0.9712849836796522, + "accuracy": 0.9712676033377647, "total_bits": 57430144, "q_proj": { "group_size": { @@ -35608,7 +35608,7 @@ } }, { - "accuracy": 0.9733225014060736, + "accuracy": 0.9733318723738194, "total_bits": 57950464, "q_proj": { "group_size": { @@ -35660,7 +35660,7 @@ } }, { - "accuracy": 0.9746621809899807, + "accuracy": 0.9746661577373743, "total_bits": 58692736, "q_proj": { "group_size": { @@ -35712,7 +35712,7 @@ } }, { - "accuracy": 0.9740584716200829, + "accuracy": 0.9740777555853128, "total_bits": 59068544, "q_proj": { "group_size": { @@ -35776,7 +35776,7 @@ } }, { - "accuracy": 0.9757704772055149, + "accuracy": 0.975718442350626, "total_bits": 59588864, "q_proj": { "group_size": { @@ -35840,7 +35840,7 @@ } }, { - "accuracy": 0.9784729983657598, + "accuracy": 0.9784317370504141, "total_bits": 61536832, "q_proj": { "group_size": { @@ -35901,7 +35901,7 @@ } }, { - "accuracy": 0.9800199028104544, + "accuracy": 0.9800386726856232, "total_bits": 62354560, "q_proj": { "group_size": { @@ -35962,7 +35962,7 @@ } }, { - "accuracy": 0.9881230108439922, + "accuracy": 0.988131619989872, "total_bits": 75246880, "q_proj": { "group_size": { @@ -36023,7 +36023,7 @@ } }, { - "accuracy": 0.9899277659133077, + "accuracy": 0.9899393441155553, "total_bits": 76510336, "q_proj": { "group_size": { @@ -36084,7 +36084,7 @@ } }, { - "accuracy": 0.9922054400667548, + "accuracy": 0.9921957161277533, "total_bits": 85667104, "q_proj": { "group_size": { @@ -36136,7 +36136,7 @@ } }, { - "accuracy": 0.9948229189030826, + "accuracy": 0.994826287496835, "total_bits": 91722880, "q_proj": { "group_size": { @@ -36188,7 +36188,7 @@ } }, { - "accuracy": 0.9979508460965008, + "accuracy": 0.9979522456414998, "total_bits": 113978656, "q_proj": { "group_size": { @@ -36242,7 +36242,7 @@ ], "model.layers.18.mlp": [ { - "accuracy": 0.8908465802669525, + "accuracy": 0.890760600566864, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -36294,7 +36294,7 @@ } }, { - "accuracy": 0.8948191478848457, + "accuracy": 0.8948594033718109, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -36346,7 +36346,7 @@ } }, { - "accuracy": 0.9099691584706306, + "accuracy": 0.9100265651941299, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -36395,7 +36395,7 @@ } }, { - "accuracy": 0.9150631278753281, + "accuracy": 0.915120542049408, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -36444,7 +36444,7 @@ } }, { - "accuracy": 0.9450795240700245, + "accuracy": 0.9450446665287018, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -36496,7 +36496,7 @@ } }, { - "accuracy": 0.949484571814537, + "accuracy": 0.9494681656360626, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -36548,7 +36548,7 @@ } }, { - "accuracy": 0.9560906924307346, + "accuracy": 0.9560775049030781, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -36597,7 +36597,7 @@ } }, { - "accuracy": 0.9707168936729431, + "accuracy": 0.9707209877669811, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -36640,7 +36640,7 @@ } }, { - "accuracy": 0.9732743687927723, + "accuracy": 0.9732760544866323, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -36683,7 +36683,7 @@ } }, { - "accuracy": 0.9719167854636908, + "accuracy": 0.9719307515770197, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -36735,7 +36735,7 @@ } }, { - "accuracy": 0.9754060637205839, + "accuracy": 0.9753952287137508, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -36787,7 +36787,7 @@ } }, { - "accuracy": 0.9856863981112838, + "accuracy": 0.9856886547058821, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -36839,7 +36839,7 @@ } }, { - "accuracy": 0.9877114491537213, + "accuracy": 0.9877165118232369, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -36891,7 +36891,7 @@ } }, { - "accuracy": 0.9919731765985489, + "accuracy": 0.9919724566861987, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -36934,7 +36934,7 @@ } }, { - "accuracy": 0.9926252705045044, + "accuracy": 0.9926214762963355, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -36983,7 +36983,7 @@ } }, { - "accuracy": 0.9940348407253623, + "accuracy": 0.9940323028713465, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -37029,7 +37029,7 @@ } }, { - "accuracy": 0.9977936870418489, + "accuracy": 0.997793102869764, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -37071,7 +37071,7 @@ ], "model.layers.19.self_attn": [ { - "accuracy": 0.9066938832402229, + "accuracy": 0.9067771062254906, "total_bits": 30308928, "q_proj": { "group_size": { @@ -37135,7 +37135,7 @@ } }, { - "accuracy": 0.9125282913446426, + "accuracy": 0.9125330448150635, "total_bits": 31455808, "q_proj": { "group_size": { @@ -37199,7 +37199,7 @@ } }, { - "accuracy": 0.9224199429154396, + "accuracy": 0.9224216490983963, "total_bits": 33412832, "q_proj": { "group_size": { @@ -37263,7 +37263,7 @@ } }, { - "accuracy": 0.942340325564146, + "accuracy": 0.9422799944877625, "total_bits": 37983200, "q_proj": { "group_size": { @@ -37327,7 +37327,7 @@ } }, { - "accuracy": 0.9528491497039795, + "accuracy": 0.9527867995202541, "total_bits": 44838176, "q_proj": { "group_size": { @@ -37391,7 +37391,7 @@ } }, { - "accuracy": 0.9537361785769463, + "accuracy": 0.9536703340709209, "total_bits": 44912768, "q_proj": { "group_size": { @@ -37455,7 +37455,7 @@ } }, { - "accuracy": 0.9727440439164639, + "accuracy": 0.9727084040641785, "total_bits": 57355552, "q_proj": { "group_size": { @@ -37507,7 +37507,7 @@ } }, { - "accuracy": 0.9736681282520294, + "accuracy": 0.9736226126551628, "total_bits": 57430144, "q_proj": { "group_size": { @@ -37559,7 +37559,7 @@ } }, { - "accuracy": 0.9758165907114744, + "accuracy": 0.9758408833295107, "total_bits": 57950464, "q_proj": { "group_size": { @@ -37611,7 +37611,7 @@ } }, { - "accuracy": 0.9769553057849407, + "accuracy": 0.976946609094739, "total_bits": 58692736, "q_proj": { "group_size": { @@ -37663,7 +37663,7 @@ } }, { - "accuracy": 0.9764603525400162, + "accuracy": 0.9764624573290348, "total_bits": 59068544, "q_proj": { "group_size": { @@ -37727,7 +37727,7 @@ } }, { - "accuracy": 0.9781377725303173, + "accuracy": 0.9780647847801447, "total_bits": 59588864, "q_proj": { "group_size": { @@ -37791,7 +37791,7 @@ } }, { - "accuracy": 0.9803321305662394, + "accuracy": 0.9802794177085161, "total_bits": 61536832, "q_proj": { "group_size": { @@ -37852,7 +37852,7 @@ } }, { - "accuracy": 0.9818277545273304, + "accuracy": 0.9817991387099028, "total_bits": 62354560, "q_proj": { "group_size": { @@ -37913,7 +37913,7 @@ } }, { - "accuracy": 0.9891065321862698, + "accuracy": 0.9891098402440548, "total_bits": 75246880, "q_proj": { "group_size": { @@ -37974,7 +37974,7 @@ } }, { - "accuracy": 0.9908470399677753, + "accuracy": 0.9908483978360891, "total_bits": 76510336, "q_proj": { "group_size": { @@ -38035,7 +38035,7 @@ } }, { - "accuracy": 0.9929368873126805, + "accuracy": 0.9929364663548768, "total_bits": 85667104, "q_proj": { "group_size": { @@ -38087,7 +38087,7 @@ } }, { - "accuracy": 0.9951926521025598, + "accuracy": 0.9951782925054431, "total_bits": 91722880, "q_proj": { "group_size": { @@ -38139,7 +38139,7 @@ } }, { - "accuracy": 0.9981394049245864, + "accuracy": 0.9981401044642553, "total_bits": 113978656, "q_proj": { "group_size": { @@ -38193,7 +38193,7 @@ ], "model.layers.19.mlp": [ { - "accuracy": 0.898843489587307, + "accuracy": 0.8988735228776932, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -38245,7 +38245,7 @@ } }, { - "accuracy": 0.9021168202161789, + "accuracy": 0.9020884037017822, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -38297,7 +38297,7 @@ } }, { - "accuracy": 0.9158534109592438, + "accuracy": 0.9158180356025696, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -38346,7 +38346,7 @@ } }, { - "accuracy": 0.9202790409326553, + "accuracy": 0.9202406033873558, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -38395,7 +38395,7 @@ } }, { - "accuracy": 0.9486076608300209, + "accuracy": 0.9486297108232975, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -38447,7 +38447,7 @@ } }, { - "accuracy": 0.9527660310268402, + "accuracy": 0.9527538195252419, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -38499,7 +38499,7 @@ } }, { - "accuracy": 0.9588682539761066, + "accuracy": 0.9588394649326801, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -38548,7 +38548,7 @@ } }, { - "accuracy": 0.9728159811347723, + "accuracy": 0.9728188179433346, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -38591,7 +38591,7 @@ } }, { - "accuracy": 0.9751972481608391, + "accuracy": 0.9751956835389137, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -38634,7 +38634,7 @@ } }, { - "accuracy": 0.9737109690904617, + "accuracy": 0.9737076684832573, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -38686,7 +38686,7 @@ } }, { - "accuracy": 0.9769691210240126, + "accuracy": 0.976981945335865, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -38738,7 +38738,7 @@ } }, { - "accuracy": 0.986573307774961, + "accuracy": 0.9865698497742414, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -38790,7 +38790,7 @@ } }, { - "accuracy": 0.9884986458346248, + "accuracy": 0.9884996181353927, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -38842,7 +38842,7 @@ } }, { - "accuracy": 0.9925676062703133, + "accuracy": 0.9925639107823372, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -38885,7 +38885,7 @@ } }, { - "accuracy": 0.993094963952899, + "accuracy": 0.993094386998564, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -38934,7 +38934,7 @@ } }, { - "accuracy": 0.9943557516671717, + "accuracy": 0.9943572375923395, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -38980,7 +38980,7 @@ } }, { - "accuracy": 0.9979684855788946, + "accuracy": 0.9979692660272121, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -39022,7 +39022,7 @@ ], "model.layers.20.self_attn": [ { - "accuracy": 0.9254191517829895, + "accuracy": 0.9252974018454552, "total_bits": 30308928, "q_proj": { "group_size": { @@ -39086,7 +39086,7 @@ } }, { - "accuracy": 0.9304513037204742, + "accuracy": 0.9304210692644119, "total_bits": 31455808, "q_proj": { "group_size": { @@ -39150,7 +39150,7 @@ } }, { - "accuracy": 0.937650989741087, + "accuracy": 0.9376390874385834, "total_bits": 33412832, "q_proj": { "group_size": { @@ -39214,7 +39214,7 @@ } }, { - "accuracy": 0.9561277441680431, + "accuracy": 0.956094067543745, "total_bits": 37983200, "q_proj": { "group_size": { @@ -39278,7 +39278,7 @@ } }, { - "accuracy": 0.9629123508930206, + "accuracy": 0.9629477187991142, "total_bits": 44838176, "q_proj": { "group_size": { @@ -39342,7 +39342,7 @@ } }, { - "accuracy": 0.9635394588112831, + "accuracy": 0.9635471068322659, "total_bits": 44912768, "q_proj": { "group_size": { @@ -39406,7 +39406,7 @@ } }, { - "accuracy": 0.9784403182566166, + "accuracy": 0.9784409739077091, "total_bits": 57355552, "q_proj": { "group_size": { @@ -39458,7 +39458,7 @@ } }, { - "accuracy": 0.9791186898946762, + "accuracy": 0.9791207388043404, "total_bits": 57430144, "q_proj": { "group_size": { @@ -39510,7 +39510,7 @@ } }, { - "accuracy": 0.9805207420140505, + "accuracy": 0.9805529490113258, "total_bits": 57950464, "q_proj": { "group_size": { @@ -39562,7 +39562,7 @@ } }, { - "accuracy": 0.981576981022954, + "accuracy": 0.9815756008028984, "total_bits": 58692736, "q_proj": { "group_size": { @@ -39614,7 +39614,7 @@ } }, { - "accuracy": 0.9814564771950245, + "accuracy": 0.981438597664237, "total_bits": 59068544, "q_proj": { "group_size": { @@ -39678,7 +39678,7 @@ } }, { - "accuracy": 0.9826366528868675, + "accuracy": 0.9826178383082151, "total_bits": 59588864, "q_proj": { "group_size": { @@ -39742,7 +39742,7 @@ } }, { - "accuracy": 0.9840204436331987, + "accuracy": 0.9840060472488403, "total_bits": 61536832, "q_proj": { "group_size": { @@ -39803,7 +39803,7 @@ } }, { - "accuracy": 0.9851696165278554, + "accuracy": 0.9851967319846153, "total_bits": 62354560, "q_proj": { "group_size": { @@ -39864,7 +39864,7 @@ } }, { - "accuracy": 0.9912118604406714, + "accuracy": 0.99119614623487, "total_bits": 75246880, "q_proj": { "group_size": { @@ -39925,7 +39925,7 @@ } }, { - "accuracy": 0.9925604872405529, + "accuracy": 0.992552753072232, "total_bits": 76510336, "q_proj": { "group_size": { @@ -39986,7 +39986,7 @@ } }, { - "accuracy": 0.9943439392372966, + "accuracy": 0.9943421361967921, "total_bits": 85667104, "q_proj": { "group_size": { @@ -40038,7 +40038,7 @@ } }, { - "accuracy": 0.9961527525447309, + "accuracy": 0.996142350602895, "total_bits": 91722880, "q_proj": { "group_size": { @@ -40090,7 +40090,7 @@ } }, { - "accuracy": 0.9985169863793999, + "accuracy": 0.9985177288763225, "total_bits": 113978656, "q_proj": { "group_size": { @@ -40144,7 +40144,7 @@ ], "model.layers.20.mlp": [ { - "accuracy": 0.9078842997550964, + "accuracy": 0.9078511148691177, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -40196,7 +40196,7 @@ } }, { - "accuracy": 0.9108857437968254, + "accuracy": 0.9108638241887093, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -40248,7 +40248,7 @@ } }, { - "accuracy": 0.9233747720718384, + "accuracy": 0.9233528077602386, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -40297,7 +40297,7 @@ } }, { - "accuracy": 0.927286222577095, + "accuracy": 0.9272778257727623, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -40346,7 +40346,7 @@ } }, { - "accuracy": 0.9533279724419117, + "accuracy": 0.9533239230513573, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -40398,7 +40398,7 @@ } }, { - "accuracy": 0.9570191614329815, + "accuracy": 0.9570236355066299, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -40450,7 +40450,7 @@ } }, { - "accuracy": 0.9625181145966053, + "accuracy": 0.962521106004715, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -40499,7 +40499,7 @@ } }, { - "accuracy": 0.9753356482833624, + "accuracy": 0.9753350578248501, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -40542,7 +40542,7 @@ } }, { - "accuracy": 0.977478351444006, + "accuracy": 0.9774747285991907, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -40585,7 +40585,7 @@ } }, { - "accuracy": 0.9761522505432367, + "accuracy": 0.9761437401175499, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -40637,7 +40637,7 @@ } }, { - "accuracy": 0.9790751449763775, + "accuracy": 0.9790640287101269, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -40689,7 +40689,7 @@ } }, { - "accuracy": 0.9878394166007638, + "accuracy": 0.9878420624881983, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -40741,7 +40741,7 @@ } }, { - "accuracy": 0.9895561756566167, + "accuracy": 0.9895521271973848, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -40793,7 +40793,7 @@ } }, { - "accuracy": 0.9932784507982433, + "accuracy": 0.993278375826776, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -40836,7 +40836,7 @@ } }, { - "accuracy": 0.9937723437324166, + "accuracy": 0.9937705183401704, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -40885,7 +40885,7 @@ } }, { - "accuracy": 0.9949068506248295, + "accuracy": 0.9949048236012459, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -40931,7 +40931,7 @@ } }, { - "accuracy": 0.9981734347529709, + "accuracy": 0.9981726757250726, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -40973,7 +40973,7 @@ ], "model.layers.21.self_attn": [ { - "accuracy": 0.9116751402616501, + "accuracy": 0.9116267189383507, "total_bits": 30308928, "q_proj": { "group_size": { @@ -41037,7 +41037,7 @@ } }, { - "accuracy": 0.9161897599697113, + "accuracy": 0.9161302745342255, "total_bits": 31455808, "q_proj": { "group_size": { @@ -41101,7 +41101,7 @@ } }, { - "accuracy": 0.9299630001187325, + "accuracy": 0.9298892468214035, "total_bits": 33412832, "q_proj": { "group_size": { @@ -41165,7 +41165,7 @@ } }, { - "accuracy": 0.951121561229229, + "accuracy": 0.951112475246191, "total_bits": 37983200, "q_proj": { "group_size": { @@ -41229,7 +41229,7 @@ } }, { - "accuracy": 0.9554766491055489, + "accuracy": 0.9555105529725552, "total_bits": 44838176, "q_proj": { "group_size": { @@ -41293,7 +41293,7 @@ } }, { - "accuracy": 0.9563397578895092, + "accuracy": 0.9563395045697689, "total_bits": 44912768, "q_proj": { "group_size": { @@ -41357,7 +41357,7 @@ } }, { - "accuracy": 0.9752995986491442, + "accuracy": 0.9753066599369049, "total_bits": 57355552, "q_proj": { "group_size": { @@ -41409,7 +41409,7 @@ } }, { - "accuracy": 0.976159306243062, + "accuracy": 0.9761861730366945, "total_bits": 57430144, "q_proj": { "group_size": { @@ -41461,7 +41461,7 @@ } }, { - "accuracy": 0.9777697827666998, + "accuracy": 0.9777861218899488, "total_bits": 57950464, "q_proj": { "group_size": { @@ -41513,7 +41513,7 @@ } }, { - "accuracy": 0.9786687269806862, + "accuracy": 0.9786933194845915, "total_bits": 58692736, "q_proj": { "group_size": { @@ -41565,7 +41565,7 @@ } }, { - "accuracy": 0.9778354410082102, + "accuracy": 0.9778481628745794, "total_bits": 59068544, "q_proj": { "group_size": { @@ -41629,7 +41629,7 @@ } }, { - "accuracy": 0.9793091975152493, + "accuracy": 0.9793014619499445, "total_bits": 59588864, "q_proj": { "group_size": { @@ -41693,7 +41693,7 @@ } }, { - "accuracy": 0.9823502618819475, + "accuracy": 0.9823220800608397, "total_bits": 61536832, "q_proj": { "group_size": { @@ -41754,7 +41754,7 @@ } }, { - "accuracy": 0.9835798889398575, + "accuracy": 0.9835791904479265, "total_bits": 62354560, "q_proj": { "group_size": { @@ -41815,7 +41815,7 @@ } }, { - "accuracy": 0.9903135802596807, + "accuracy": 0.9903201386332512, "total_bits": 75246880, "q_proj": { "group_size": { @@ -41876,7 +41876,7 @@ } }, { - "accuracy": 0.9917551334947348, + "accuracy": 0.9917648062109947, "total_bits": 76510336, "q_proj": { "group_size": { @@ -41937,7 +41937,7 @@ } }, { - "accuracy": 0.993569896556437, + "accuracy": 0.9935718071646988, "total_bits": 85667104, "q_proj": { "group_size": { @@ -41989,7 +41989,7 @@ } }, { - "accuracy": 0.9958970271982253, + "accuracy": 0.995899414177984, "total_bits": 91722880, "q_proj": { "group_size": { @@ -42041,7 +42041,7 @@ } }, { - "accuracy": 0.9983291744720191, + "accuracy": 0.9983296216232702, "total_bits": 113978656, "q_proj": { "group_size": { @@ -42095,7 +42095,7 @@ ], "model.layers.21.mlp": [ { - "accuracy": 0.9132195189595222, + "accuracy": 0.9132231697440147, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -42147,7 +42147,7 @@ } }, { - "accuracy": 0.9158970862627029, + "accuracy": 0.9158701747655869, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -42199,7 +42199,7 @@ } }, { - "accuracy": 0.9271287769079208, + "accuracy": 0.9270192459225655, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -42248,7 +42248,7 @@ } }, { - "accuracy": 0.9307196065783501, + "accuracy": 0.9306086674332619, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -42297,7 +42297,7 @@ } }, { - "accuracy": 0.955941203981638, + "accuracy": 0.9559643566608429, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -42349,7 +42349,7 @@ } }, { - "accuracy": 0.9593969620764256, + "accuracy": 0.9593847468495369, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -42401,7 +42401,7 @@ } }, { - "accuracy": 0.9644332267343998, + "accuracy": 0.9644359201192856, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -42450,7 +42450,7 @@ } }, { - "accuracy": 0.9767589289695024, + "accuracy": 0.9767424874007702, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -42493,7 +42493,7 @@ } }, { - "accuracy": 0.9787649121135473, + "accuracy": 0.9787469636648893, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -42536,7 +42536,7 @@ } }, { - "accuracy": 0.9774891547858715, + "accuracy": 0.9774995297193527, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -42588,7 +42588,7 @@ } }, { - "accuracy": 0.9802252352237701, + "accuracy": 0.9802324343472719, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -42640,7 +42640,7 @@ } }, { - "accuracy": 0.9885327704250813, + "accuracy": 0.9885260602459311, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -42692,7 +42692,7 @@ } }, { - "accuracy": 0.9901311350986362, + "accuracy": 0.9901291783899069, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -42744,7 +42744,7 @@ } }, { - "accuracy": 0.9936433634720743, + "accuracy": 0.9936425597406924, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -42787,7 +42787,7 @@ } }, { - "accuracy": 0.9941183044575155, + "accuracy": 0.9941207375377417, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -42836,7 +42836,7 @@ } }, { - "accuracy": 0.995133682154119, + "accuracy": 0.9951349799521267, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -42882,7 +42882,7 @@ } }, { - "accuracy": 0.9982258901000023, + "accuracy": 0.9982257036026567, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -42924,7 +42924,7 @@ ], "model.layers.22.self_attn": [ { - "accuracy": 0.8600220680236816, + "accuracy": 0.8604224324226379, "total_bits": 30308928, "q_proj": { "group_size": { @@ -42988,7 +42988,7 @@ } }, { - "accuracy": 0.8723062872886658, + "accuracy": 0.8720817565917969, "total_bits": 31455808, "q_proj": { "group_size": { @@ -43052,7 +43052,7 @@ } }, { - "accuracy": 0.8845950290560722, + "accuracy": 0.8842579200863838, "total_bits": 33412832, "q_proj": { "group_size": { @@ -43116,7 +43116,7 @@ } }, { - "accuracy": 0.9110095426440239, + "accuracy": 0.9107557758688927, "total_bits": 37983200, "q_proj": { "group_size": { @@ -43180,7 +43180,7 @@ } }, { - "accuracy": 0.9318739250302315, + "accuracy": 0.9317922592163086, "total_bits": 44838176, "q_proj": { "group_size": { @@ -43244,7 +43244,7 @@ } }, { - "accuracy": 0.9331164434552193, + "accuracy": 0.933014988899231, "total_bits": 44912768, "q_proj": { "group_size": { @@ -43308,7 +43308,7 @@ } }, { - "accuracy": 0.9576672576367855, + "accuracy": 0.9576846696436405, "total_bits": 57355552, "q_proj": { "group_size": { @@ -43360,7 +43360,7 @@ } }, { - "accuracy": 0.9591643884778023, + "accuracy": 0.9591707922518253, "total_bits": 57430144, "q_proj": { "group_size": { @@ -43412,7 +43412,7 @@ } }, { - "accuracy": 0.9620198979973793, + "accuracy": 0.9620674923062325, "total_bits": 57950464, "q_proj": { "group_size": { @@ -43464,7 +43464,7 @@ } }, { - "accuracy": 0.9642005227506161, + "accuracy": 0.9641453549265862, "total_bits": 58692736, "q_proj": { "group_size": { @@ -43516,7 +43516,7 @@ } }, { - "accuracy": 0.9660037606954575, + "accuracy": 0.9659963771700859, "total_bits": 59068544, "q_proj": { "group_size": { @@ -43580,7 +43580,7 @@ } }, { - "accuracy": 0.968350138515234, + "accuracy": 0.9683596417307854, "total_bits": 59588864, "q_proj": { "group_size": { @@ -43644,7 +43644,7 @@ } }, { - "accuracy": 0.9704683814197779, + "accuracy": 0.970446715131402, "total_bits": 61536832, "q_proj": { "group_size": { @@ -43705,7 +43705,7 @@ } }, { - "accuracy": 0.9727780558168888, + "accuracy": 0.9728066977113485, "total_bits": 62354560, "q_proj": { "group_size": { @@ -43766,7 +43766,7 @@ } }, { - "accuracy": 0.9837654158473015, + "accuracy": 0.9837965164333582, "total_bits": 75246880, "q_proj": { "group_size": { @@ -43827,7 +43827,7 @@ } }, { - "accuracy": 0.9864024138078094, + "accuracy": 0.9864012431353331, "total_bits": 76510336, "q_proj": { "group_size": { @@ -43888,7 +43888,7 @@ } }, { - "accuracy": 0.988879844546318, + "accuracy": 0.9888947391882539, "total_bits": 85667104, "q_proj": { "group_size": { @@ -43940,7 +43940,7 @@ } }, { - "accuracy": 0.9924624958075583, + "accuracy": 0.9924732316285372, "total_bits": 91722880, "q_proj": { "group_size": { @@ -43992,7 +43992,7 @@ } }, { - "accuracy": 0.9970627035945654, + "accuracy": 0.997059682616964, "total_bits": 113978656, "q_proj": { "group_size": { @@ -44046,7 +44046,7 @@ ], "model.layers.22.mlp": [ { - "accuracy": 0.9158392548561096, + "accuracy": 0.9158530607819557, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -44098,7 +44098,7 @@ } }, { - "accuracy": 0.9184657260775566, + "accuracy": 0.9185132682323456, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -44150,7 +44150,7 @@ } }, { - "accuracy": 0.9284176975488663, + "accuracy": 0.9284360781311989, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -44199,7 +44199,7 @@ } }, { - "accuracy": 0.9315355196595192, + "accuracy": 0.9315476194024086, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -44248,7 +44248,7 @@ } }, { - "accuracy": 0.9572760388255119, + "accuracy": 0.9572666324675083, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -44300,7 +44300,7 @@ } }, { - "accuracy": 0.9605343677103519, + "accuracy": 0.9605594612658024, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -44352,7 +44352,7 @@ } }, { - "accuracy": 0.9649710729718208, + "accuracy": 0.9650028869509697, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -44401,7 +44401,7 @@ } }, { - "accuracy": 0.9774835146963596, + "accuracy": 0.9774914421141148, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -44444,7 +44444,7 @@ } }, { - "accuracy": 0.9793780688196421, + "accuracy": 0.9793844223022461, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -44487,7 +44487,7 @@ } }, { - "accuracy": 0.9781624637544155, + "accuracy": 0.9781443998217583, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -44539,7 +44539,7 @@ } }, { - "accuracy": 0.9807386361062527, + "accuracy": 0.9807199090719223, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -44591,7 +44591,7 @@ } }, { - "accuracy": 0.988841949030757, + "accuracy": 0.9888424472883344, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -44643,7 +44643,7 @@ } }, { - "accuracy": 0.9903620686382055, + "accuracy": 0.9903717981651425, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -44695,7 +44695,7 @@ } }, { - "accuracy": 0.9938594889827073, + "accuracy": 0.9938576300628483, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -44738,7 +44738,7 @@ } }, { - "accuracy": 0.9942582775838673, + "accuracy": 0.994258991908282, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -44787,7 +44787,7 @@ } }, { - "accuracy": 0.9951435467228293, + "accuracy": 0.9951435327529907, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -44833,7 +44833,7 @@ } }, { - "accuracy": 0.9983168488834053, + "accuracy": 0.9983171256026253, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -44875,7 +44875,7 @@ ], "model.layers.23.self_attn": [ { - "accuracy": 0.9002235606312752, + "accuracy": 0.9000980034470558, "total_bits": 30308928, "q_proj": { "group_size": { @@ -44939,7 +44939,7 @@ } }, { - "accuracy": 0.9065486118197441, + "accuracy": 0.9066225737333298, "total_bits": 31455808, "q_proj": { "group_size": { @@ -45003,7 +45003,7 @@ } }, { - "accuracy": 0.918041318655014, + "accuracy": 0.9180461913347244, "total_bits": 33412832, "q_proj": { "group_size": { @@ -45067,7 +45067,7 @@ } }, { - "accuracy": 0.9384324066340923, + "accuracy": 0.9384226053953171, "total_bits": 37983200, "q_proj": { "group_size": { @@ -45131,7 +45131,7 @@ } }, { - "accuracy": 0.9503814466297626, + "accuracy": 0.950372826308012, "total_bits": 44838176, "q_proj": { "group_size": { @@ -45195,7 +45195,7 @@ } }, { - "accuracy": 0.9512348845601082, + "accuracy": 0.9512467607855797, "total_bits": 44912768, "q_proj": { "group_size": { @@ -45259,7 +45259,7 @@ } }, { - "accuracy": 0.970819503068924, + "accuracy": 0.9708252772688866, "total_bits": 57355552, "q_proj": { "group_size": { @@ -45311,7 +45311,7 @@ } }, { - "accuracy": 0.9717874713242054, + "accuracy": 0.9717806540429592, "total_bits": 57430144, "q_proj": { "group_size": { @@ -45363,7 +45363,7 @@ } }, { - "accuracy": 0.9738185089081526, + "accuracy": 0.9737685434520245, "total_bits": 57950464, "q_proj": { "group_size": { @@ -45415,7 +45415,7 @@ } }, { - "accuracy": 0.9751602187752724, + "accuracy": 0.9751146398484707, "total_bits": 58692736, "q_proj": { "group_size": { @@ -45467,7 +45467,7 @@ } }, { - "accuracy": 0.9751874227076769, + "accuracy": 0.9751594942063093, "total_bits": 59068544, "q_proj": { "group_size": { @@ -45531,7 +45531,7 @@ } }, { - "accuracy": 0.9768790081143379, + "accuracy": 0.9768443964421749, "total_bits": 59588864, "q_proj": { "group_size": { @@ -45595,7 +45595,7 @@ } }, { - "accuracy": 0.9792286064475775, + "accuracy": 0.9791899565607309, "total_bits": 61536832, "q_proj": { "group_size": { @@ -45656,7 +45656,7 @@ } }, { - "accuracy": 0.9807712472975254, + "accuracy": 0.9807248823344707, "total_bits": 62354560, "q_proj": { "group_size": { @@ -45717,7 +45717,7 @@ } }, { - "accuracy": 0.9885837156325579, + "accuracy": 0.98859893810004, "total_bits": 75246880, "q_proj": { "group_size": { @@ -45778,7 +45778,7 @@ } }, { - "accuracy": 0.9903419725596905, + "accuracy": 0.9903517542406917, "total_bits": 76510336, "q_proj": { "group_size": { @@ -45839,7 +45839,7 @@ } }, { - "accuracy": 0.9923813547939062, + "accuracy": 0.9923766339197755, "total_bits": 85667104, "q_proj": { "group_size": { @@ -45891,7 +45891,7 @@ } }, { - "accuracy": 0.9948961278423667, + "accuracy": 0.9948966847732663, "total_bits": 91722880, "q_proj": { "group_size": { @@ -45943,7 +45943,7 @@ } }, { - "accuracy": 0.9980002944357693, + "accuracy": 0.9980048832949251, "total_bits": 113978656, "q_proj": { "group_size": { @@ -45997,7 +45997,7 @@ ], "model.layers.23.mlp": [ { - "accuracy": 0.9183400496840477, + "accuracy": 0.9182481989264488, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -46049,7 +46049,7 @@ } }, { - "accuracy": 0.9207602441310883, + "accuracy": 0.9207983389496803, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -46101,7 +46101,7 @@ } }, { - "accuracy": 0.9304101318120956, + "accuracy": 0.9304468557238579, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -46150,7 +46150,7 @@ } }, { - "accuracy": 0.9335396587848663, + "accuracy": 0.9335753992199898, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -46199,7 +46199,7 @@ } }, { - "accuracy": 0.95809031650424, + "accuracy": 0.9580657295882702, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -46251,7 +46251,7 @@ } }, { - "accuracy": 0.9615879841148853, + "accuracy": 0.9615712836384773, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -46303,7 +46303,7 @@ } }, { - "accuracy": 0.9659944400191307, + "accuracy": 0.9659555852413177, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -46352,7 +46352,7 @@ } }, { - "accuracy": 0.9778957311064005, + "accuracy": 0.9778776671737432, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -46395,7 +46395,7 @@ } }, { - "accuracy": 0.9797942880541086, + "accuracy": 0.9797797352075577, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -46438,7 +46438,7 @@ } }, { - "accuracy": 0.9784904830157757, + "accuracy": 0.9784805551171303, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -46490,7 +46490,7 @@ } }, { - "accuracy": 0.9812441393733025, + "accuracy": 0.9812456294894218, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -46542,7 +46542,7 @@ } }, { - "accuracy": 0.9889877140522003, + "accuracy": 0.9889850988984108, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -46594,7 +46594,7 @@ } }, { - "accuracy": 0.9906198102980852, + "accuracy": 0.9906222112476826, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -46646,7 +46646,7 @@ } }, { - "accuracy": 0.9938806621357799, + "accuracy": 0.9938766132108867, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -46689,7 +46689,7 @@ } }, { - "accuracy": 0.9943293896503747, + "accuracy": 0.9943245043978095, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -46738,7 +46738,7 @@ } }, { - "accuracy": 0.9952116911299527, + "accuracy": 0.9952051844447851, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -46784,7 +46784,7 @@ } }, { - "accuracy": 0.9983016916085035, + "accuracy": 0.9983023661188781, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -46826,7 +46826,7 @@ ], "model.layers.24.self_attn": [ { - "accuracy": 0.9000411629676819, + "accuracy": 0.9000758603215218, "total_bits": 30308928, "q_proj": { "group_size": { @@ -46890,7 +46890,7 @@ } }, { - "accuracy": 0.9092481806874275, + "accuracy": 0.9089899435639381, "total_bits": 31455808, "q_proj": { "group_size": { @@ -46954,7 +46954,7 @@ } }, { - "accuracy": 0.9228631407022476, + "accuracy": 0.9227519929409027, "total_bits": 33412832, "q_proj": { "group_size": { @@ -47018,7 +47018,7 @@ } }, { - "accuracy": 0.9451075606048107, + "accuracy": 0.9450107850134373, "total_bits": 37983200, "q_proj": { "group_size": { @@ -47082,7 +47082,7 @@ } }, { - "accuracy": 0.9511439241468906, + "accuracy": 0.9511034078896046, "total_bits": 44838176, "q_proj": { "group_size": { @@ -47146,7 +47146,7 @@ } }, { - "accuracy": 0.9521462395787239, + "accuracy": 0.9521391801536083, "total_bits": 44912768, "q_proj": { "group_size": { @@ -47210,7 +47210,7 @@ } }, { - "accuracy": 0.9704149011522532, + "accuracy": 0.9703994449228048, "total_bits": 57355552, "q_proj": { "group_size": { @@ -47262,7 +47262,7 @@ } }, { - "accuracy": 0.9715412594377995, + "accuracy": 0.9715345222502947, "total_bits": 57430144, "q_proj": { "group_size": { @@ -47314,7 +47314,7 @@ } }, { - "accuracy": 0.9736700374633074, + "accuracy": 0.9736871141940355, "total_bits": 57950464, "q_proj": { "group_size": { @@ -47366,7 +47366,7 @@ } }, { - "accuracy": 0.9749877713620663, + "accuracy": 0.9749988541007042, "total_bits": 58692736, "q_proj": { "group_size": { @@ -47418,7 +47418,7 @@ } }, { - "accuracy": 0.9757188335061073, + "accuracy": 0.9757179487496614, "total_bits": 59068544, "q_proj": { "group_size": { @@ -47482,7 +47482,7 @@ } }, { - "accuracy": 0.9774507116526365, + "accuracy": 0.9774703308939934, "total_bits": 59588864, "q_proj": { "group_size": { @@ -47546,7 +47546,7 @@ } }, { - "accuracy": 0.9802383426576853, + "accuracy": 0.9802435729652643, "total_bits": 61536832, "q_proj": { "group_size": { @@ -47607,7 +47607,7 @@ } }, { - "accuracy": 0.9814821984618902, + "accuracy": 0.9814784899353981, "total_bits": 62354560, "q_proj": { "group_size": { @@ -47668,7 +47668,7 @@ } }, { - "accuracy": 0.9890236798673868, + "accuracy": 0.9890194535255432, "total_bits": 75246880, "q_proj": { "group_size": { @@ -47729,7 +47729,7 @@ } }, { - "accuracy": 0.9906842615455389, + "accuracy": 0.9906802931800485, "total_bits": 76510336, "q_proj": { "group_size": { @@ -47790,7 +47790,7 @@ } }, { - "accuracy": 0.9923012899234891, + "accuracy": 0.992301972117275, "total_bits": 85667104, "q_proj": { "group_size": { @@ -47842,7 +47842,7 @@ } }, { - "accuracy": 0.9948786222375929, + "accuracy": 0.9948805985040963, "total_bits": 91722880, "q_proj": { "group_size": { @@ -47894,7 +47894,7 @@ } }, { - "accuracy": 0.997995373327285, + "accuracy": 0.9979949889238924, "total_bits": 113978656, "q_proj": { "group_size": { @@ -47948,7 +47948,7 @@ ], "model.layers.24.mlp": [ { - "accuracy": 0.92559964209795, + "accuracy": 0.9256472215056419, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -48000,7 +48000,7 @@ } }, { - "accuracy": 0.9278130382299423, + "accuracy": 0.9277902320027351, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -48052,7 +48052,7 @@ } }, { - "accuracy": 0.9358445852994919, + "accuracy": 0.9358237609267235, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -48101,7 +48101,7 @@ } }, { - "accuracy": 0.9386767745018005, + "accuracy": 0.9386411532759666, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -48150,7 +48150,7 @@ } }, { - "accuracy": 0.9618233405053616, + "accuracy": 0.9618184603750706, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -48202,7 +48202,7 @@ } }, { - "accuracy": 0.9650960303843021, + "accuracy": 0.9650743305683136, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -48254,7 +48254,7 @@ } }, { - "accuracy": 0.9686915799975395, + "accuracy": 0.9686741307377815, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -48303,7 +48303,7 @@ } }, { - "accuracy": 0.9795693270862103, + "accuracy": 0.9795593917369843, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -48346,7 +48346,7 @@ } }, { - "accuracy": 0.9815245009958744, + "accuracy": 0.9815181270241737, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -48389,7 +48389,7 @@ } }, { - "accuracy": 0.9804772697389126, + "accuracy": 0.9804782513529062, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -48441,7 +48441,7 @@ } }, { - "accuracy": 0.982961056753993, + "accuracy": 0.9829648770391941, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -48493,7 +48493,7 @@ } }, { - "accuracy": 0.9900317844003439, + "accuracy": 0.9900405630469322, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -48545,7 +48545,7 @@ } }, { - "accuracy": 0.9914947487413883, + "accuracy": 0.9914911538362503, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -48597,7 +48597,7 @@ } }, { - "accuracy": 0.994380182120949, + "accuracy": 0.9943781793117523, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -48640,7 +48640,7 @@ } }, { - "accuracy": 0.9948879033327103, + "accuracy": 0.9948839037679136, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -48689,7 +48689,7 @@ } }, { - "accuracy": 0.9956484977155924, + "accuracy": 0.9956428701989353, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -48735,7 +48735,7 @@ } }, { - "accuracy": 0.9983874682802707, + "accuracy": 0.9983870283467695, "total_bits": 512046144, "gate_proj": { "group_size": { @@ -48777,7 +48777,7 @@ ], "model.layers.25.self_attn": [ { - "accuracy": 0.9237724244594574, + "accuracy": 0.923765018582344, "total_bits": 30308928, "q_proj": { "group_size": { @@ -48841,7 +48841,7 @@ } }, { - "accuracy": 0.9273674637079239, + "accuracy": 0.9273531660437584, "total_bits": 31455808, "q_proj": { "group_size": { @@ -48905,7 +48905,7 @@ } }, { - "accuracy": 0.9392730295658112, + "accuracy": 0.9392255060374737, "total_bits": 33412832, "q_proj": { "group_size": { @@ -48969,7 +48969,7 @@ } }, { - "accuracy": 0.9596496224403381, + "accuracy": 0.9596178978681564, "total_bits": 37983200, "q_proj": { "group_size": { @@ -49033,7 +49033,7 @@ } }, { - "accuracy": 0.9613975808024406, + "accuracy": 0.9613914974033833, "total_bits": 44838176, "q_proj": { "group_size": { @@ -49097,7 +49097,7 @@ } }, { - "accuracy": 0.9620046243071556, + "accuracy": 0.962019756436348, "total_bits": 44912768, "q_proj": { "group_size": { @@ -49161,7 +49161,7 @@ } }, { - "accuracy": 0.9788154270499945, + "accuracy": 0.9787868969142437, "total_bits": 57355552, "q_proj": { "group_size": { @@ -49213,7 +49213,7 @@ } }, { - "accuracy": 0.9793564602732658, + "accuracy": 0.9793211314827204, "total_bits": 57430144, "q_proj": { "group_size": { @@ -49265,7 +49265,7 @@ } }, { - "accuracy": 0.980896295979619, + "accuracy": 0.9808873273432255, "total_bits": 57950464, "q_proj": { "group_size": { @@ -49317,7 +49317,7 @@ } }, { - "accuracy": 0.9818288143724203, + "accuracy": 0.9818263165652752, "total_bits": 58692736, "q_proj": { "group_size": { @@ -49369,7 +49369,7 @@ } }, { - "accuracy": 0.9807142727077007, + "accuracy": 0.9806932359933853, "total_bits": 59068544, "q_proj": { "group_size": { @@ -49433,7 +49433,7 @@ } }, { - "accuracy": 0.9820475969463587, + "accuracy": 0.9820628892630339, "total_bits": 59588864, "q_proj": { "group_size": { @@ -49497,7 +49497,7 @@ } }, { - "accuracy": 0.9845238020643592, + "accuracy": 0.9845330221578479, "total_bits": 61536832, "q_proj": { "group_size": { @@ -49558,7 +49558,7 @@ } }, { - "accuracy": 0.9857827695086598, + "accuracy": 0.9857928091660142, "total_bits": 62354560, "q_proj": { "group_size": { @@ -49619,7 +49619,7 @@ } }, { - "accuracy": 0.9914994724094868, + "accuracy": 0.9915020493790507, "total_bits": 75246880, "q_proj": { "group_size": { @@ -49680,7 +49680,7 @@ } }, { - "accuracy": 0.9928717431612313, + "accuracy": 0.9928759732283652, "total_bits": 76510336, "q_proj": { "group_size": { @@ -49741,7 +49741,7 @@ } }, { - "accuracy": 0.9944902677088976, + "accuracy": 0.9944881168194115, "total_bits": 85667104, "q_proj": { "group_size": { @@ -49793,7 +49793,7 @@ } }, { - "accuracy": 0.9964965777471662, + "accuracy": 0.9964962496887892, "total_bits": 91722880, "q_proj": { "group_size": { @@ -49845,7 +49845,7 @@ } }, { - "accuracy": 0.9985775835812092, + "accuracy": 0.998579174396582, "total_bits": 113978656, "q_proj": { "group_size": { @@ -49899,7 +49899,7 @@ ], "model.layers.25.mlp": [ { - "accuracy": 0.9394260719418526, + "accuracy": 0.9394301772117615, "total_bits": 142969344, "gate_proj": { "group_size": { @@ -49951,7 +49951,7 @@ } }, { - "accuracy": 0.9412130787968636, + "accuracy": 0.9411459490656853, "total_bits": 148277760, "gate_proj": { "group_size": { @@ -50003,7 +50003,7 @@ } }, { - "accuracy": 0.9476935565471649, + "accuracy": 0.9477050378918648, "total_bits": 165462912, "gate_proj": { "group_size": { @@ -50052,7 +50052,7 @@ } }, { - "accuracy": 0.9500809907913208, + "accuracy": 0.9500880539417267, "total_bits": 185590656, "gate_proj": { "group_size": { @@ -50101,7 +50101,7 @@ } }, { - "accuracy": 0.9688832182437181, + "accuracy": 0.96889528632164, "total_bits": 209798784, "gate_proj": { "group_size": { @@ -50153,7 +50153,7 @@ } }, { - "accuracy": 0.9714504927396774, + "accuracy": 0.9714055731892586, "total_bits": 214864128, "gate_proj": { "group_size": { @@ -50205,7 +50205,7 @@ } }, { - "accuracy": 0.9744708389043808, + "accuracy": 0.9744217358529568, "total_bits": 230963904, "gate_proj": { "group_size": { @@ -50254,7 +50254,7 @@ } }, { - "accuracy": 0.9832967892289162, + "accuracy": 0.9833166711032391, "total_bits": 263770272, "gate_proj": { "group_size": { @@ -50297,7 +50297,7 @@ } }, { - "accuracy": 0.9849021304398775, + "accuracy": 0.9849346242845058, "total_bits": 267653376, "gate_proj": { "group_size": { @@ -50340,7 +50340,7 @@ } }, { - "accuracy": 0.9840464647859335, + "accuracy": 0.9840444251894951, "total_bits": 272393856, "gate_proj": { "group_size": { @@ -50392,7 +50392,7 @@ } }, { - "accuracy": 0.9860300803557038, + "accuracy": 0.9860321069136262, "total_bits": 277459200, "gate_proj": { "group_size": { @@ -50444,7 +50444,7 @@ } }, { - "accuracy": 0.9918431174010038, + "accuracy": 0.9918465595692396, "total_bits": 334988928, "gate_proj": { "group_size": { @@ -50496,7 +50496,7 @@ } }, { - "accuracy": 0.9930016789585352, + "accuracy": 0.9929989366792142, "total_bits": 340054272, "gate_proj": { "group_size": { @@ -50548,7 +50548,7 @@ } }, { - "accuracy": 0.9954412854276597, + "accuracy": 0.9954434814862907, "total_bits": 386967168, "gate_proj": { "group_size": { @@ -50591,7 +50591,7 @@ } }, { - "accuracy": 0.9958284390158951, + "accuracy": 0.9958340055309236, "total_bits": 400569408, "gate_proj": { "group_size": { @@ -50640,7 +50640,7 @@ } }, { - "accuracy": 0.9965037242509425, + "accuracy": 0.9965090290643275, "total_bits": 436548672, "gate_proj": { "group_size": { @@ -50686,7 +50686,7 @@ } }, { - "accuracy": 0.9987250744597986, + "accuracy": 0.9987249359255657, "total_bits": 512046144, "gate_proj": { "group_size": { diff --git a/out_tensor/lm_head.safetensors b/out_tensor/lm_head.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..124b8c2b2c29a5c8e32d4a9f905833ababcfaad0 --- /dev/null +++ b/out_tensor/lm_head.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b2bc1ad72140f40d2d73acb37acddc4704cc5358f0f378caa10f0e58541045fc +size 469257756 diff --git a/out_tensor/model.layers.0.mlp.down_proj.safetensors b/out_tensor/model.layers.0.mlp.down_proj.safetensors index a10a8532cfc99f8a14c58caaac8fbc92e4995813..2ea67208b512e4b74b1b842100e387a93a216d37 100644 --- a/out_tensor/model.layers.0.mlp.down_proj.safetensors +++ b/out_tensor/model.layers.0.mlp.down_proj.safetensors @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:d2edb50847068d249fd17b57ebea37ee42bb56763ba9efb56c0f8cd954320301 -size 11807960 +oid sha256:b5404282df3d7167b31c5ce12f26e5b262e98f2f656775e22c613d5b90aecb66 +size 11540696 diff --git a/out_tensor/model.layers.0.mlp.gate_proj.safetensors b/out_tensor/model.layers.0.mlp.gate_proj.safetensors index b4c7910e2a28a1f84626ddc0d7b70600482d3652..420879886bc13f5bd66b33f3c84322360b38b561 100644 --- a/out_tensor/model.layers.0.mlp.gate_proj.safetensors +++ b/out_tensor/model.layers.0.mlp.gate_proj.safetensors @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:003009306f40d1854edb77e44c621688ec96051cc499178f2c0cea95c85a75d1 -size 11253696 +oid sha256:9779e16bdc1e3f79c9b1dff697cf7a61613f50a8fe284e1411f16483d4ac7283 +size 10958784 diff --git a/out_tensor/model.layers.0.mlp.up_proj.safetensors b/out_tensor/model.layers.0.mlp.up_proj.safetensors index 329db4842e242a2f8fa787ea9d80dbde934f8827..1430d26e121c25f5855508291e708cc8c3d70e8c 100644 --- a/out_tensor/model.layers.0.mlp.up_proj.safetensors +++ b/out_tensor/model.layers.0.mlp.up_proj.safetensors @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:d833214f664d36476c9f4ac26ab0e60418e6dba54a6e3824632005db1ca591ce -size 11622328 +oid sha256:97d8aab8f35eb7ad82df59e9b3923f9599bcd39968b01b7f2deae76c57f8277e +size 10958776 diff --git a/out_tensor/model.layers.0.self_attn.o_proj.safetensors b/out_tensor/model.layers.0.self_attn.o_proj.safetensors index dc7f894e651568ee7279e0766b2293d4da0353f8..9b0c688844a99105c242384f2f651307c409c3f9 100644 --- a/out_tensor/model.layers.0.self_attn.o_proj.safetensors +++ b/out_tensor/model.layers.0.self_attn.o_proj.safetensors @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:3601d3ef6fda80a4e54a78eb30ee989ce3ca68524a4de5cf91f138e01cd01bfe +oid sha256:c2db05a8562a2cc236829b78c1adf68bd67487dae5b6fc9f4336061ad2824168 size 2478808 diff --git a/out_tensor/model.layers.0.self_attn.q_proj.safetensors b/out_tensor/model.layers.0.self_attn.q_proj.safetensors index f5c81d49606ce6fa5c85192cf583e4f6eff9d652..acfcaead32a36f3dd2b89a2a0eff2e7f0d304128 100644 --- a/out_tensor/model.layers.0.self_attn.q_proj.safetensors +++ b/out_tensor/model.layers.0.self_attn.q_proj.safetensors @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:490cd38cb9e0af07b8a7424005ffa7df7fcfd972fea25587f8878e0d0088f56b +oid sha256:9ea89129704afdcc98be029d39c45e04b546dcdb74d9b698ccf88d92f293f065 size 2471664 diff --git a/out_tensor/model.layers.0.self_attn.v_proj.safetensors b/out_tensor/model.layers.0.self_attn.v_proj.safetensors index a66d11eebfb60043a84553ff33243ea346968de2..ae7e228c8ba6a91e69ac51c2f9b90ebc2acfbd1f 100644 --- a/out_tensor/model.layers.0.self_attn.v_proj.safetensors +++ b/out_tensor/model.layers.0.self_attn.v_proj.safetensors @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:9dae9865b4f598ffb36cdcdd102b1b3d641b32f3d77c9c48a9d6928035f04280 +oid sha256:735ca65ca23a30ea68792b8914f7ca90987104d13a2cb8792d1ebb37af7fb826 size 1259464 diff --git a/out_tensor/model.layers.1.mlp.down_proj.safetensors b/out_tensor/model.layers.1.mlp.down_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..7b8daf86ef2815253178ad58a96cd9b89379a115 --- /dev/null +++ b/out_tensor/model.layers.1.mlp.down_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:019900b1046bc0e5bafb46e3f7f2afa4df9919c75ce631a8dd6978531e35d2f0 +size 11807960 diff --git a/out_tensor/model.layers.1.mlp.gate_proj.safetensors b/out_tensor/model.layers.1.mlp.gate_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..d05934beff88a09ff7832c6bf820876511ae0b10 --- /dev/null +++ b/out_tensor/model.layers.1.mlp.gate_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:83237b2740518ed5c5f30b2f38b3fb944c5d89726e56dc16eb308304cbff462b +size 11253696 diff --git a/out_tensor/model.layers.1.mlp.up_proj.safetensors b/out_tensor/model.layers.1.mlp.up_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..a22d708fe31c4933ac7583218300d4c576ac8d85 --- /dev/null +++ b/out_tensor/model.layers.1.mlp.up_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:256d05cd3a39e13f8be838305fe02ccc59d22b75270f447c3d51644c6c750711 +size 11622328 diff --git a/out_tensor/model.layers.1.self_attn.k_proj.safetensors b/out_tensor/model.layers.1.self_attn.k_proj.safetensors index 2567cbf711f698ad59222876fd8836b2cc161007..12f9a8accc1676e7b60f9f52c9b7c1fe0e411b66 100644 --- a/out_tensor/model.layers.1.self_attn.k_proj.safetensors +++ b/out_tensor/model.layers.1.self_attn.k_proj.safetensors @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:4784ffd2c379e18e2d802a82920bc8ec16f3218b134777edc4c16985322bb515 -size 1788548 +oid sha256:44637721beca8f3ccc5d84eb25405bd8503a16e813af0481d17e19b40ebe076c +size 1526404 diff --git a/out_tensor/model.layers.1.self_attn.o_proj.safetensors b/out_tensor/model.layers.1.self_attn.o_proj.safetensors index 3995193d96cf3d9da3088e5fb5a8dde416e9ea13..c2cbbf7701663f4eff7483ea3ea97075a7a6f07e 100644 --- a/out_tensor/model.layers.1.self_attn.o_proj.safetensors +++ b/out_tensor/model.layers.1.self_attn.o_proj.safetensors @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:558e1de0bac39b09867bfb5286291a83689c3e2655fbcf3b859330573675908d -size 3566200 +oid sha256:b971f2d12ade22eef6ef0ad2a5ab26e0bc75e217b02bdb6323cb5737c64a4ce6 +size 3050104 diff --git a/out_tensor/model.layers.1.self_attn.q_proj.safetensors b/out_tensor/model.layers.1.self_attn.q_proj.safetensors index 51b157acd8df998073980f1b5e000385365d349a..0fd555acebd98a9d74ef1cab5de3337f43d0fc56 100644 --- a/out_tensor/model.layers.1.self_attn.q_proj.safetensors +++ b/out_tensor/model.layers.1.self_attn.q_proj.safetensors @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:d398f5db36ea429f7377a4c60e7c1ca67dc3b9a55d337ff5f97b39b0ce590968 -size 3567236 +oid sha256:ba57b69f71bac6d37f328fd15934627cd1e10f3d1b17bc98f72459da3bca960b +size 3042948 diff --git a/out_tensor/model.layers.1.self_attn.v_proj.safetensors b/out_tensor/model.layers.1.self_attn.v_proj.safetensors index ee10bad5a5cdf4516bfffd480067876d50fccd7e..14e98498f1da3c4a2dfeed6b3508d8f31f88d693 100644 --- a/out_tensor/model.layers.1.self_attn.v_proj.safetensors +++ b/out_tensor/model.layers.1.self_attn.v_proj.safetensors @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:e3ba91db45cab6520c19d0113e85374d61de030fd8a49957ee4f6c90b172936c +oid sha256:51c5fc96c5d1b2b830e4d68f5bafcecfb37b57887276df0b4d165ad31d3ae1d9 size 1788548 diff --git a/out_tensor/model.layers.10.mlp.down_proj.safetensors b/out_tensor/model.layers.10.mlp.down_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..803328706a4c4e6e8355b9482d7e9f998cdf8a6f --- /dev/null +++ b/out_tensor/model.layers.10.mlp.down_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0fd32a645b8b0f12e3df2c132250098f4ffc648c2bded362d05c31e874d88321 +size 11540696 diff --git a/out_tensor/model.layers.10.mlp.gate_proj.safetensors b/out_tensor/model.layers.10.mlp.gate_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..5555bf42d80daa3121fc7ad3875305cc08995b50 --- /dev/null +++ b/out_tensor/model.layers.10.mlp.gate_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ca2d530b682c8e757dcc56b619ee8ccb462972624a5902eb48710d66a7ae06f8 +size 10958792 diff --git a/out_tensor/model.layers.10.mlp.up_proj.safetensors b/out_tensor/model.layers.10.mlp.up_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..88e7aa150a2b82491bf19a805f96d50d0ea70597 --- /dev/null +++ b/out_tensor/model.layers.10.mlp.up_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2e887f6591ea1096bb1faecf4dbd0eb0a87c7bfe7c7c9c5059992b23d418e39e +size 10958784 diff --git a/out_tensor/model.layers.10.self_attn.k_proj.safetensors b/out_tensor/model.layers.10.self_attn.k_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..d3e441cfaba3def381ae87ce5c46d74346f018d4 --- /dev/null +++ b/out_tensor/model.layers.10.self_attn.k_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e9dda2e6776875327462523bc48d73343bb42279252703c39bfdecfd9574a68e +size 1240824 diff --git a/out_tensor/model.layers.10.self_attn.o_proj.safetensors b/out_tensor/model.layers.10.self_attn.o_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..ae98172a8ecc8e13699cf813ba2b4e5482cd3ded --- /dev/null +++ b/out_tensor/model.layers.10.self_attn.o_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:89c781cc5448ff3c9c21798744206c561b2f2b9624183a18ff5d47216624e4b1 +size 2478816 diff --git a/out_tensor/model.layers.10.self_attn.q_proj.safetensors b/out_tensor/model.layers.10.self_attn.q_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3b14894be1ce85fef6833d35ff65724b6ce29808 --- /dev/null +++ b/out_tensor/model.layers.10.self_attn.q_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5c14808cb3dee4abeb8f9d8aa953df16b068982309beb4ca9a752d286d4c4cb6 +size 2471672 diff --git a/out_tensor/model.layers.10.self_attn.v_proj.safetensors b/out_tensor/model.layers.10.self_attn.v_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..282ee5cddc82e5bdb3922e3e4306b1f62cb161a5 --- /dev/null +++ b/out_tensor/model.layers.10.self_attn.v_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bd035353287769134bf370d2989ce6ef0a8a20d309687d2f92de13b452342a9f +size 1502968 diff --git a/out_tensor/model.layers.11.mlp.down_proj.safetensors b/out_tensor/model.layers.11.mlp.down_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..74b593c8752515e094a40cb83210865b48f2af41 --- /dev/null +++ b/out_tensor/model.layers.11.mlp.down_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1f800986ce7b64311b441f9bd9f74391338c2783184fd4e232115fa178dca367 +size 11304464 diff --git a/out_tensor/model.layers.11.mlp.gate_proj.safetensors b/out_tensor/model.layers.11.mlp.gate_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..dcb8ac8662978164e76cdf0f20f9e947287691e7 --- /dev/null +++ b/out_tensor/model.layers.11.mlp.gate_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fb3f544cf86e06127d1fdcfd41481a3dfbe834513219fd9c4d2c270fafdbc6b4 +size 10709636 diff --git a/out_tensor/model.layers.11.mlp.up_proj.safetensors b/out_tensor/model.layers.11.mlp.up_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..bda91fe3bd73c0f5179feb863bfb2539828368ea --- /dev/null +++ b/out_tensor/model.layers.11.mlp.up_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4b04bc434794c446aa3468f1a98ef074e65be6906953bcc999fdaef5c99d07c4 +size 10958784 diff --git a/out_tensor/model.layers.11.self_attn.k_proj.safetensors b/out_tensor/model.layers.11.self_attn.k_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..de183d5fbbf4390892ee65b6a669df66725ac3a6 --- /dev/null +++ b/out_tensor/model.layers.11.self_attn.k_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d0eeb89074f3d314e91974cdd44ed0f6ddeb74e86ed285e9282a8a6650bd1afa +size 1208056 diff --git a/out_tensor/model.layers.11.self_attn.o_proj.safetensors b/out_tensor/model.layers.11.self_attn.o_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..a5fbdb50c6d02ef763dded42e4e944b0b4b9891e --- /dev/null +++ b/out_tensor/model.layers.11.self_attn.o_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7ddffca85d4e1c9e1df82bbe2213a785ec50fd97f0e9b0bbb998ee3f43294d4d +size 2405088 diff --git a/out_tensor/model.layers.11.self_attn.q_proj.safetensors b/out_tensor/model.layers.11.self_attn.q_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..ecd18e458829387b486f678f24d308c129b7269c --- /dev/null +++ b/out_tensor/model.layers.11.self_attn.q_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:29b3ecdc0b88278afd0fd7ebedccb3ed187887cc1badd8d1c2a0d54cb38bb400 +size 2406136 diff --git a/out_tensor/model.layers.11.self_attn.v_proj.safetensors b/out_tensor/model.layers.11.self_attn.v_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..943843cbad93e7c08d3c369b379835354a1ab3bc --- /dev/null +++ b/out_tensor/model.layers.11.self_attn.v_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cbb4ba7cf48d0c740329434e7c73a230b184d40f54aee1b0cb6dc92a1e6fbeb0 +size 1226704 diff --git a/out_tensor/model.layers.12.mlp.down_proj.safetensors b/out_tensor/model.layers.12.mlp.down_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..9dd1593faf11e4a4fd9544326d2d602505ec521e --- /dev/null +++ b/out_tensor/model.layers.12.mlp.down_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c4c6577ea81a2f256fa8fe08181da735ede95ef5d6b74e0f0d2d6cad568b72c7 +size 11540696 diff --git a/out_tensor/model.layers.12.mlp.gate_proj.safetensors b/out_tensor/model.layers.12.mlp.gate_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..6e884883a379b85304a69ac05c2b4d07da6955a8 --- /dev/null +++ b/out_tensor/model.layers.12.mlp.gate_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:818333870b3bf7d6e4c1e59282471581737a49b838a6703701abe6cfac36f58b +size 10958792 diff --git a/out_tensor/model.layers.12.mlp.up_proj.safetensors b/out_tensor/model.layers.12.mlp.up_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..b5d4bcbb57edea01c1f38d5a08f00ccdf4229b02 --- /dev/null +++ b/out_tensor/model.layers.12.mlp.up_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1968ea6475c7809b9573227ed85c52a6bcbe3902c9893ece7da931efe990b2cc +size 10958784 diff --git a/out_tensor/model.layers.12.self_attn.k_proj.safetensors b/out_tensor/model.layers.12.self_attn.k_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..04ccf4775c3e8f7688dfb5a8f0bd7a77359d2a45 --- /dev/null +++ b/out_tensor/model.layers.12.self_attn.k_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:90b245c1046d9082eb3fe68edfd64efcd4ea583ae8765c90ccfca2048e053eba +size 1788556 diff --git a/out_tensor/model.layers.12.self_attn.o_proj.safetensors b/out_tensor/model.layers.12.self_attn.o_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..a5a13b3426ff11cf2e93bd65d2bc27bbe6ca7132 --- /dev/null +++ b/out_tensor/model.layers.12.self_attn.o_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cc878a6f7d91c6b84b2b61b7e50648cee82e68303b719b84786482392f77c666 +size 3566208 diff --git a/out_tensor/model.layers.12.self_attn.q_proj.safetensors b/out_tensor/model.layers.12.self_attn.q_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..40bdb59a4d08657c490a529a794145afb64768b7 --- /dev/null +++ b/out_tensor/model.layers.12.self_attn.q_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d20b78b5d0555964dfd3e8d5e7101ff1ce2df301b8b38bed5048b29f622177db +size 3567244 diff --git a/out_tensor/model.layers.12.self_attn.v_proj.safetensors b/out_tensor/model.layers.12.self_attn.v_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..57f8152ecd04e5c8a8dd65a020cb2f9e879030cc --- /dev/null +++ b/out_tensor/model.layers.12.self_attn.v_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:45fae4540b7f74d2ffa68a794f68a736ee0e340de30a7181ba182234600ca357 +size 1788556 diff --git a/out_tensor/model.layers.13.mlp.down_proj.safetensors b/out_tensor/model.layers.13.mlp.down_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..a9f10bd1b7e6e8bc219d051a89b48aee863800d9 --- /dev/null +++ b/out_tensor/model.layers.13.mlp.down_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6b7c69102f3af7a03752bea91b07c5afd0898c3f6b332896738176f51ebbc893 +size 11540696 diff --git a/out_tensor/model.layers.13.mlp.gate_proj.safetensors b/out_tensor/model.layers.13.mlp.gate_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..12e7fe995797064bd145816b30925dd3df2eec6b --- /dev/null +++ b/out_tensor/model.layers.13.mlp.gate_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1746030b041ab25d8f1419a78f0270b75e28f30067cd4726a2468aa899a981b5 +size 10958792 diff --git a/out_tensor/model.layers.13.mlp.up_proj.safetensors b/out_tensor/model.layers.13.mlp.up_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..05c19e373012ad042a5d9536cd68f717ad1ca6c2 --- /dev/null +++ b/out_tensor/model.layers.13.mlp.up_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e34143ade090b979db74fbfbeca57b5171fce1882c6b2750bd4e407bdcf1385a +size 10958784 diff --git a/out_tensor/model.layers.13.self_attn.k_proj.safetensors b/out_tensor/model.layers.13.self_attn.k_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..e4ee1e891d741ed1eb8e7559b4501893029e8b3f --- /dev/null +++ b/out_tensor/model.layers.13.self_attn.k_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:733ccf754b38a2c3f8f7be975350a88a8173c4de7082ee973b93bcdccb5b1904 +size 1554384 diff --git a/out_tensor/model.layers.13.self_attn.o_proj.safetensors b/out_tensor/model.layers.13.self_attn.o_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..5ece90a2d4db08fbb815bb6337e4e4d2fdd542e2 --- /dev/null +++ b/out_tensor/model.layers.13.self_attn.o_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8fe1c55ebb9efa9193c51ddb45f6642b79c1fc4ca6139e62ca86788148aa026d +size 3096480 diff --git a/out_tensor/model.layers.13.self_attn.q_proj.safetensors b/out_tensor/model.layers.13.self_attn.q_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..88fe1cd73ef3587f59d1633c8d0d8fbdefe5e3fa --- /dev/null +++ b/out_tensor/model.layers.13.self_attn.q_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c3385285b239166553244a1a7046e675fc9e9331e742ae8d6adeae4ae5704bf6 +size 3098576 diff --git a/out_tensor/model.layers.13.self_attn.v_proj.safetensors b/out_tensor/model.layers.13.self_attn.v_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3eb3bf5df703d550292ee62147db5c835bca8865 --- /dev/null +++ b/out_tensor/model.layers.13.self_attn.v_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fef6613abd140d956ecfa19c0afb33202cb116c4da109e6b92bb39bef9570f49 +size 1816528 diff --git a/out_tensor/model.layers.14.mlp.down_proj.safetensors b/out_tensor/model.layers.14.mlp.down_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..f68688283ba9948affc0a4aeb0f2b32341899919 --- /dev/null +++ b/out_tensor/model.layers.14.mlp.down_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ca35b0255cb9f7a571b67686bcedc84758e3afe6ac3fa330f49fd92293ad0a04 +size 11540696 diff --git a/out_tensor/model.layers.14.mlp.gate_proj.safetensors b/out_tensor/model.layers.14.mlp.gate_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..8cf9e7203978e6ed4cdc3dd5cf9b75370c4b31a9 --- /dev/null +++ b/out_tensor/model.layers.14.mlp.gate_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cc2fa38f60829d5e89b7beaccaff8508303bae05838efbcbb2dac9476ef041d5 +size 10958792 diff --git a/out_tensor/model.layers.14.mlp.up_proj.safetensors b/out_tensor/model.layers.14.mlp.up_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..04f0ddc9755a369fed3316b56d8e4e9a9ebd3a11 --- /dev/null +++ b/out_tensor/model.layers.14.mlp.up_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4f8c84ba547fd90493696f800e703cb00980b74fe3c18cba1f609aedf1d38e1d +size 10958784 diff --git a/out_tensor/model.layers.14.self_attn.k_proj.safetensors b/out_tensor/model.layers.14.self_attn.k_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..040f0f8d6adc1b63b5a59d5dd5eabcd018111cff --- /dev/null +++ b/out_tensor/model.layers.14.self_attn.k_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8c922226f6a0b9883bfd71654d03e28f4f82dc636b17ce80e83c9fac22a85386 +size 1554384 diff --git a/out_tensor/model.layers.14.self_attn.o_proj.safetensors b/out_tensor/model.layers.14.self_attn.o_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..e35b86c69d104296a04e7aabff5acbf0da1aad72 --- /dev/null +++ b/out_tensor/model.layers.14.self_attn.o_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f20784d45bb17704d4ec9073337a04953e00bb94e742edf46ecd82b86276bbf7 +size 3096480 diff --git a/out_tensor/model.layers.14.self_attn.q_proj.safetensors b/out_tensor/model.layers.14.self_attn.q_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..c27165a25a0c599962e99e44f7b0f646921324f8 --- /dev/null +++ b/out_tensor/model.layers.14.self_attn.q_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c40e91cfb000ef93cfcc5e671e0092d470a4a6f3c48be49754a0d6afededddfc +size 3098576 diff --git a/out_tensor/model.layers.14.self_attn.v_proj.safetensors b/out_tensor/model.layers.14.self_attn.v_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..64e638b80cccf0db006f08c6d504c8a4b58c9537 --- /dev/null +++ b/out_tensor/model.layers.14.self_attn.v_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4227af9e9a5ed8037cee6289ac420dc5e6377d89d4ca44a41324970769788afc +size 1816528 diff --git a/out_tensor/model.layers.15.mlp.down_proj.safetensors b/out_tensor/model.layers.15.mlp.down_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..d0655b38126ccc41b70571b7bb424a44df968151 --- /dev/null +++ b/out_tensor/model.layers.15.mlp.down_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7438de03ffa6c17e83f58bc6208fb5d4a3bacc5438750d2a7bf97da39ea901ce +size 11304464 diff --git a/out_tensor/model.layers.15.mlp.gate_proj.safetensors b/out_tensor/model.layers.15.mlp.gate_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..4718e894753b2d881e2d7b3ebd1fe4354120c887 --- /dev/null +++ b/out_tensor/model.layers.15.mlp.gate_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4ba967cac6901f0948da158e709d0155d4e42a5b3298a0e6f12fe7438b739db9 +size 10709636 diff --git a/out_tensor/model.layers.15.mlp.up_proj.safetensors b/out_tensor/model.layers.15.mlp.up_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..a08896d1b002cefba20bb3b020d8a2454d7b058d --- /dev/null +++ b/out_tensor/model.layers.15.mlp.up_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:01b9cdf2f21d13fa8df95423c02dd616c19023ac03e1534207ebd7583cf4cfd2 +size 10958784 diff --git a/out_tensor/model.layers.15.self_attn.k_proj.safetensors b/out_tensor/model.layers.15.self_attn.k_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..9633a1da43bdcbda01c6cdec1b6e75b908e9d626 --- /dev/null +++ b/out_tensor/model.layers.15.self_attn.k_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5935e52e09a95cee1dbe1467b6a66bb8c761bf6e3037ddd125e43824691de242 +size 1554384 diff --git a/out_tensor/model.layers.15.self_attn.o_proj.safetensors b/out_tensor/model.layers.15.self_attn.o_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..4a325ebe66d1dfe2aa4b105873b79b0711539fe7 --- /dev/null +++ b/out_tensor/model.layers.15.self_attn.o_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f94603551b2e6457f300270b21a10d9110972dfb0b9610cccd197974a3116bf2 +size 3096480 diff --git a/out_tensor/model.layers.15.self_attn.q_proj.safetensors b/out_tensor/model.layers.15.self_attn.q_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..1e238cfec4a02b9bb7b7a838eb2cc827686dbb6a --- /dev/null +++ b/out_tensor/model.layers.15.self_attn.q_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:579932c3ed5532d96e5ba0f230c3774f8c95c16e5f4d39123ae0816b3131fd24 +size 3098576 diff --git a/out_tensor/model.layers.15.self_attn.v_proj.safetensors b/out_tensor/model.layers.15.self_attn.v_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..c75913c49d23ed7108008212bdc4bb31fe5eb33f --- /dev/null +++ b/out_tensor/model.layers.15.self_attn.v_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a98735ae2ec225ea7b291f3d53736f9be9daed5f74bad66e0d298b58a1b183a9 +size 1816528 diff --git a/out_tensor/model.layers.16.mlp.down_proj.safetensors b/out_tensor/model.layers.16.mlp.down_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..ada1d61622d6e5c382bbb7f052b7e271a91bd192 --- /dev/null +++ b/out_tensor/model.layers.16.mlp.down_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:beb240dac300b1628e1b956a79ba5146b1d4a9e4086236e1a8c55395ccf422c5 +size 11807960 diff --git a/out_tensor/model.layers.16.mlp.gate_proj.safetensors b/out_tensor/model.layers.16.mlp.gate_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..b5b2e208fb8a436886f73dcd2c59158eaac97db1 --- /dev/null +++ b/out_tensor/model.layers.16.mlp.gate_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6bfe583b25301db52ee62d48f1658d536f2ba60c1d789880656dc23c797bd072 +size 11253704 diff --git a/out_tensor/model.layers.16.mlp.up_proj.safetensors b/out_tensor/model.layers.16.mlp.up_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..a468ed8403c5ebd444a4bec54e219f2df6787212 --- /dev/null +++ b/out_tensor/model.layers.16.mlp.up_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3acf9435f91d02f8232cc56e07d85b485e5ad4ec2d657621489330d29a359c4e +size 11622336 diff --git a/out_tensor/model.layers.16.self_attn.k_proj.safetensors b/out_tensor/model.layers.16.self_attn.k_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..c594a571f0748e57d1e7bec5fd3a6f5a0cea4f6b --- /dev/null +++ b/out_tensor/model.layers.16.self_attn.k_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2a324c1d6de640905035095cc4e23eb8c79c71d0686e5488c2082b7146ac009e +size 1554384 diff --git a/out_tensor/model.layers.16.self_attn.o_proj.safetensors b/out_tensor/model.layers.16.self_attn.o_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..8d9a297bff04f217ecd87a7fce9119d27b5ebff9 --- /dev/null +++ b/out_tensor/model.layers.16.self_attn.o_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4241c28ff26256052871369fcd5c411ac0b0888f869a8f07da2de1a92382460a +size 3096480 diff --git a/out_tensor/model.layers.16.self_attn.q_proj.safetensors b/out_tensor/model.layers.16.self_attn.q_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..a65c260af0ccb9afa65f9958b726d5ea3d9a49d8 --- /dev/null +++ b/out_tensor/model.layers.16.self_attn.q_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:25494f6b6143eb90e63fe1cb66f064042bf046fdea01f7c6c3ab91b7efc3a43a +size 3098576 diff --git a/out_tensor/model.layers.16.self_attn.v_proj.safetensors b/out_tensor/model.layers.16.self_attn.v_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..8d4afe3cb39b045ba8f16a11c3a8d6dbee2b1a95 --- /dev/null +++ b/out_tensor/model.layers.16.self_attn.v_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3d2a4c20190756ec15b1e2b6aa3564fac925091b555be47ec53d71fa115e98e7 +size 1816528 diff --git a/out_tensor/model.layers.17.mlp.down_proj.safetensors b/out_tensor/model.layers.17.mlp.down_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..5d0c2e9746b28558cc4317b1269358633ba4ebfa --- /dev/null +++ b/out_tensor/model.layers.17.mlp.down_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f74ed0bedb582028051eab6e0d0bd6e7f43f62fb116fd2b874911591364ae1d0 +size 11540696 diff --git a/out_tensor/model.layers.17.mlp.gate_proj.safetensors b/out_tensor/model.layers.17.mlp.gate_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..86b54c8466adf1c43a10c382a7ae7e312dd89e2b --- /dev/null +++ b/out_tensor/model.layers.17.mlp.gate_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:debb3c3e7cf146b3973bbae476cdc2bd9645db1aa36182b852d698820c8aa2d6 +size 10958792 diff --git a/out_tensor/model.layers.17.mlp.up_proj.safetensors b/out_tensor/model.layers.17.mlp.up_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..813dfc9558d529aaca532f762a2c07510d8931bd --- /dev/null +++ b/out_tensor/model.layers.17.mlp.up_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:471c08595edf9c0755ba6560b3b238f46c9f6b5cf46c6b14fc7820dd49beadda +size 10958784 diff --git a/out_tensor/model.layers.17.self_attn.k_proj.safetensors b/out_tensor/model.layers.17.self_attn.k_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..529721729d39e3b10ab433e7f8a7d17dcbe58a05 --- /dev/null +++ b/out_tensor/model.layers.17.self_attn.k_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1caa61eda49dafdbb0f2cc6fd7465b0b02f882121f2e50b762afea295609433c +size 1554384 diff --git a/out_tensor/model.layers.17.self_attn.o_proj.safetensors b/out_tensor/model.layers.17.self_attn.o_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..d920413281cd6b5d67c16b9c99d9fc60ee227ba4 --- /dev/null +++ b/out_tensor/model.layers.17.self_attn.o_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dc381f17b56fa13087b197bb75d164bbc50502711f260b425309e3c6fded68c2 +size 3096480 diff --git a/out_tensor/model.layers.17.self_attn.q_proj.safetensors b/out_tensor/model.layers.17.self_attn.q_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..6132c983402407ca64bd540b098a9a7f50b81ae7 --- /dev/null +++ b/out_tensor/model.layers.17.self_attn.q_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4a309ec2072f7cb987926d6df94dd1ee8bd3cecc4177c4ec5269f97869292460 +size 3098576 diff --git a/out_tensor/model.layers.17.self_attn.v_proj.safetensors b/out_tensor/model.layers.17.self_attn.v_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..61f2e04efe178d2b51c4470940fc635d569f1b6e --- /dev/null +++ b/out_tensor/model.layers.17.self_attn.v_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6fc3edb9fd37489bd5cd9a28f87d5aab7c809d23cddfd1938f9e30cc6385c491 +size 1816528 diff --git a/out_tensor/model.layers.18.mlp.down_proj.safetensors b/out_tensor/model.layers.18.mlp.down_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..5fae8502a244cfc0c4080dd1426c94d5272e26a2 --- /dev/null +++ b/out_tensor/model.layers.18.mlp.down_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2e612e0c2b8b4b723cc84f91dfc7be6e4dff2e5acbfdbc2e9f596f85dcbf89b2 +size 11540696 diff --git a/out_tensor/model.layers.18.mlp.gate_proj.safetensors b/out_tensor/model.layers.18.mlp.gate_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..55fb34ce85efa9c22a46c16f1509645f74a03cbc --- /dev/null +++ b/out_tensor/model.layers.18.mlp.gate_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7a111478d7d158c79a9097ffeaba3675c3b04ae81d4eb05c53cce5fb39dfd6cc +size 10958792 diff --git a/out_tensor/model.layers.18.mlp.up_proj.safetensors b/out_tensor/model.layers.18.mlp.up_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..5d684396daf780a7ab6ead59912ab209c7075ea1 --- /dev/null +++ b/out_tensor/model.layers.18.mlp.up_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4b53ce6d0ba4ae1cdce9c121a27ee679a4dc4bc97afa2c1a3db85bdc391eff56 +size 10958784 diff --git a/out_tensor/model.layers.18.self_attn.k_proj.safetensors b/out_tensor/model.layers.18.self_attn.k_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..1078d82a696470d7467829d1d08063312d79c1f6 --- /dev/null +++ b/out_tensor/model.layers.18.self_attn.k_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0416a5d742417095c7ebe00a56eb3b11e5f9e49e242fd78795d0a3a41ea171bb +size 1554384 diff --git a/out_tensor/model.layers.18.self_attn.o_proj.safetensors b/out_tensor/model.layers.18.self_attn.o_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..93f8e746b9a9cb42ee18065cb3dc7e89e53ebee4 --- /dev/null +++ b/out_tensor/model.layers.18.self_attn.o_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7476d362567be845a3c370b96da5698f6ac5a851eb4cc13df5a3ebbf0e859e3a +size 3096480 diff --git a/out_tensor/model.layers.18.self_attn.q_proj.safetensors b/out_tensor/model.layers.18.self_attn.q_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..c8c283eaf6dff8128cacba37a22d1ac9220b1920 --- /dev/null +++ b/out_tensor/model.layers.18.self_attn.q_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:396b16307cef8f5dec4dffb0c06b5cc6008ea17f658b21c329bab19ad1bc43c3 +size 3098576 diff --git a/out_tensor/model.layers.18.self_attn.v_proj.safetensors b/out_tensor/model.layers.18.self_attn.v_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..8b19a3e765988448a3406e28dd1dd51b26c66338 --- /dev/null +++ b/out_tensor/model.layers.18.self_attn.v_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d3f55aff56d6228eb625c7825bc05365e2e666e882cb1ebccdc9f4a39ceb34cd +size 1816528 diff --git a/out_tensor/model.layers.19.mlp.down_proj.safetensors b/out_tensor/model.layers.19.mlp.down_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..d382da46e30b5a5ba6f4790d68a2faa6572c44bd --- /dev/null +++ b/out_tensor/model.layers.19.mlp.down_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0c0ae762a0455d42ce080a803cc96504c99cb0087c5e4a9a96b6dc0262f14c66 +size 11540696 diff --git a/out_tensor/model.layers.19.mlp.gate_proj.safetensors b/out_tensor/model.layers.19.mlp.gate_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..0083c03b0367b332002ced58d817d0db02297eed --- /dev/null +++ b/out_tensor/model.layers.19.mlp.gate_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:10c9b6ba04677c0156b6ce6440237c90fb22d1073d80ddb09d9cf0aa94ff594d +size 10958792 diff --git a/out_tensor/model.layers.19.mlp.up_proj.safetensors b/out_tensor/model.layers.19.mlp.up_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..c58e4baf98c668ca657e09877b94c1c43ea3caa9 --- /dev/null +++ b/out_tensor/model.layers.19.mlp.up_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:29bad2ed4ff31497a16e75a7104255ca1ce06fc2ecc2bcc1570ceefe25da911d +size 10958784 diff --git a/out_tensor/model.layers.19.self_attn.k_proj.safetensors b/out_tensor/model.layers.19.self_attn.k_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..db55492ed4c682bc8aae94d789fb39fff2a415e9 --- /dev/null +++ b/out_tensor/model.layers.19.self_attn.k_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:29be287d614e7631d0b6a9cf4981f62ddbea9e7e3debe7e5487f38443c836184 +size 1259472 diff --git a/out_tensor/model.layers.19.self_attn.o_proj.safetensors b/out_tensor/model.layers.19.self_attn.o_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..182798fdd4cf0c2c0bed39cc94df9cf99bab7a61 --- /dev/null +++ b/out_tensor/model.layers.19.self_attn.o_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5ae695af3907144c4d5d85362903b660b85df6b952b826b524b72df9c012352b +size 2506656 diff --git a/out_tensor/model.layers.19.self_attn.q_proj.safetensors b/out_tensor/model.layers.19.self_attn.q_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..e7489d3c67bcd66b188868535b2f6473b5cc070b --- /dev/null +++ b/out_tensor/model.layers.19.self_attn.q_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:80009e0e37c27ea6724ab376d2fe24051652402051636d17bbf37b63c6ede56e +size 2508752 diff --git a/out_tensor/model.layers.19.self_attn.v_proj.safetensors b/out_tensor/model.layers.19.self_attn.v_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..ee3f57a64bdb723132063591effbc76aefe7cfcc --- /dev/null +++ b/out_tensor/model.layers.19.self_attn.v_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a9bec3687712ed777477a63830c4c8b1131b6a15aa4818134668548adb466b30 +size 1521616 diff --git a/out_tensor/model.layers.2.mlp.down_proj.safetensors b/out_tensor/model.layers.2.mlp.down_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..a69061c884c940e52daf69276c7170fcf0429b91 --- /dev/null +++ b/out_tensor/model.layers.2.mlp.down_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8834eb5195c21cfe425255a511ff3fe09ad4ba2734329c3f17a02b6cf209062d +size 11807960 diff --git a/out_tensor/model.layers.2.mlp.gate_proj.safetensors b/out_tensor/model.layers.2.mlp.gate_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..4eff3bfcc751c95fd95a8c3db380b5906ce8bc6d --- /dev/null +++ b/out_tensor/model.layers.2.mlp.gate_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6ca9d38c20da195cca647f4e25d5c1f8be4bf850335afdb4015edfa781ff9146 +size 11253696 diff --git a/out_tensor/model.layers.2.mlp.up_proj.safetensors b/out_tensor/model.layers.2.mlp.up_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..0db97d87c28289d4281c9aa132d033795a66a09d --- /dev/null +++ b/out_tensor/model.layers.2.mlp.up_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3c627cbf5ca2937c4b7b7ce9fcececa9ca791d4f058d811fbb22bd5c18f06b13 +size 11622328 diff --git a/out_tensor/model.layers.2.self_attn.k_proj.safetensors b/out_tensor/model.layers.2.self_attn.k_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..b6100fe454fd3b002eb2d37585e36d76828ba8dd --- /dev/null +++ b/out_tensor/model.layers.2.self_attn.k_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0cb5cd6767e1e86117302b8656fee4976633895fb0490cd61c72c93aac2ee5b6 +size 1526404 diff --git a/out_tensor/model.layers.2.self_attn.o_proj.safetensors b/out_tensor/model.layers.2.self_attn.o_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..d3c485f334f2c4ad2dc92f898ebd8c74b7e359d5 --- /dev/null +++ b/out_tensor/model.layers.2.self_attn.o_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b0d6aac1312a44f997145613262c567bb5fa779af4b8b4fa8a243d502ffa1b11 +size 3050104 diff --git a/out_tensor/model.layers.2.self_attn.q_proj.safetensors b/out_tensor/model.layers.2.self_attn.q_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..9244eb7a8921924048bdf77d49ff787b34b93b11 --- /dev/null +++ b/out_tensor/model.layers.2.self_attn.q_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5858ac145405f7e500bd8e2eccd63ab27a3d40e3b6975639900195aaa3e68566 +size 3042948 diff --git a/out_tensor/model.layers.2.self_attn.v_proj.safetensors b/out_tensor/model.layers.2.self_attn.v_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..977079b1f0bb40f445a6268d8c05b594eb782980 --- /dev/null +++ b/out_tensor/model.layers.2.self_attn.v_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4ebd2636282ea19e6a96729f077f7004a540096782dad4fae3b0bac694645718 +size 1788548 diff --git a/out_tensor/model.layers.20.mlp.down_proj.safetensors b/out_tensor/model.layers.20.mlp.down_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..2d955acf0e87ebe29feef156e5c290e0d3a31f20 --- /dev/null +++ b/out_tensor/model.layers.20.mlp.down_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6196758001ddd1c2afbb80e12be5bdd26f92a57b4dfb04ed559a1ee8bb2bc788 +size 11540696 diff --git a/out_tensor/model.layers.20.mlp.gate_proj.safetensors b/out_tensor/model.layers.20.mlp.gate_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..deed3e74d06c1242fa172be343c0619c828d9cfe --- /dev/null +++ b/out_tensor/model.layers.20.mlp.gate_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c4e09a68d2c42a72d2bd559e548ad500027e3be0dff3fe9f7cefb5a4e385f362 +size 10958792 diff --git a/out_tensor/model.layers.20.mlp.up_proj.safetensors b/out_tensor/model.layers.20.mlp.up_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..c076472e42d82401f84130c5ec5352b9da35ec24 --- /dev/null +++ b/out_tensor/model.layers.20.mlp.up_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:35c166992f0bbf4300b71bbbc11becf25cc526cc70814c1aae8837ac825de336 +size 10958784 diff --git a/out_tensor/model.layers.20.self_attn.k_proj.safetensors b/out_tensor/model.layers.20.self_attn.k_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..c652814c23393928bcfc99dda26f01a40f6e241e --- /dev/null +++ b/out_tensor/model.layers.20.self_attn.k_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7f08df970e7ef3412f179d02652469fcc7eb3003aec9df350fa3122ecc96710c +size 1816528 diff --git a/out_tensor/model.layers.20.self_attn.o_proj.safetensors b/out_tensor/model.layers.20.self_attn.o_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..161cf7adb7ae56ff48c28fcb7ae8a4ffe8e9de8e --- /dev/null +++ b/out_tensor/model.layers.20.self_attn.o_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:930f0fc13f689917c37c1f9610881b25bebae7434447efd7300f7ac3bb629e35 +size 3621792 diff --git a/out_tensor/model.layers.20.self_attn.q_proj.safetensors b/out_tensor/model.layers.20.self_attn.q_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..320016443501411f4eae7dd891e434e6622eddc4 --- /dev/null +++ b/out_tensor/model.layers.20.self_attn.q_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:81be5292ca10cc2575c46ffa2f38ab007bfb18a11800b49c4ec07183c1cedb37 +size 3622864 diff --git a/out_tensor/model.layers.20.self_attn.v_proj.safetensors b/out_tensor/model.layers.20.self_attn.v_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..fe1c77e871955dd21e87df38dc6791241329bcfa --- /dev/null +++ b/out_tensor/model.layers.20.self_attn.v_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7052f0a3c77e190fd5a112f1c359e0da52d6c013c45c88a355d3fe203a823b03 +size 2406352 diff --git a/out_tensor/model.layers.21.mlp.down_proj.safetensors b/out_tensor/model.layers.21.mlp.down_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..2e0fd3d307227f41a5ceadbd9b0e68c5e605415c --- /dev/null +++ b/out_tensor/model.layers.21.mlp.down_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e6862999ee1af574a9c5fc16eba523181c525413098e6704f1c6a2de49966e4b +size 11540696 diff --git a/out_tensor/model.layers.21.mlp.gate_proj.safetensors b/out_tensor/model.layers.21.mlp.gate_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..cfa7c3f017420bf1c2a67c080eadab2651ba3bf7 --- /dev/null +++ b/out_tensor/model.layers.21.mlp.gate_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:93198272d4998975c7ac7826aca3c8627cb51105027c00d8486f62278e414a41 +size 10958792 diff --git a/out_tensor/model.layers.21.mlp.up_proj.safetensors b/out_tensor/model.layers.21.mlp.up_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..13841bab932716e76d27ef747d8a07b8164a368d --- /dev/null +++ b/out_tensor/model.layers.21.mlp.up_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:42170b4b82e8c51250a95602b9c4f646216e0cc67c68253e1cedf0120b3b2f9b +size 10958784 diff --git a/out_tensor/model.layers.21.self_attn.k_proj.safetensors b/out_tensor/model.layers.21.self_attn.k_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..f401d0ad9219f03859cf670b6e4738e58f6f504f --- /dev/null +++ b/out_tensor/model.layers.21.self_attn.k_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:694652ddfd09f8f981bebd7a480d1774e5773abc3c8bcc8a49f59a8345a7d9a4 +size 1816528 diff --git a/out_tensor/model.layers.21.self_attn.o_proj.safetensors b/out_tensor/model.layers.21.self_attn.o_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..e62c15362c7961b235c8e1635ffcedbe240531cb --- /dev/null +++ b/out_tensor/model.layers.21.self_attn.o_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:28ce095044c06a8e118cda664d2aa7c1e57146025ab3d804c6aec266f69f32da +size 3621792 diff --git a/out_tensor/model.layers.21.self_attn.q_proj.safetensors b/out_tensor/model.layers.21.self_attn.q_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..037a025f5d84a7d339e81689ce45506e10babd05 --- /dev/null +++ b/out_tensor/model.layers.21.self_attn.q_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:349d118c074708e3d7be47dbe0332bfb0194aef6dfcb22a0955c97f6922b14d0 +size 3622864 diff --git a/out_tensor/model.layers.21.self_attn.v_proj.safetensors b/out_tensor/model.layers.21.self_attn.v_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..c9273b652b6bcab5d6e5b21e93bfd09118b0fa19 --- /dev/null +++ b/out_tensor/model.layers.21.self_attn.v_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:37f1dd068db997dd67cf2f00d44ca55ad691e30db23129bd7f8086fb39b66f8d +size 2406352 diff --git a/out_tensor/model.layers.22.mlp.down_proj.safetensors b/out_tensor/model.layers.22.mlp.down_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..cc953092aa080df9890090a5eb468ef9b1078a22 --- /dev/null +++ b/out_tensor/model.layers.22.mlp.down_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7367b8f20b23e018ac7d041411be13ad862b4523632a06c66da0132e31c13fdf +size 11807960 diff --git a/out_tensor/model.layers.22.mlp.gate_proj.safetensors b/out_tensor/model.layers.22.mlp.gate_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..1971c0fe66690c33765cabb704f21fbcb042e870 --- /dev/null +++ b/out_tensor/model.layers.22.mlp.gate_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c7714754fdfb9e37fa43b12aa96785571c49dc1ba1419a0496344e98b94d502d +size 11253704 diff --git a/out_tensor/model.layers.22.mlp.up_proj.safetensors b/out_tensor/model.layers.22.mlp.up_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..d0bb877cc5fddf0e76f3117c6fb8604e4e8e53b4 --- /dev/null +++ b/out_tensor/model.layers.22.mlp.up_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:322c31ed2c5bf12c182acb12f332f70051e139fe0e054ae847d3189e45bbe518 +size 11622336 diff --git a/out_tensor/model.layers.22.self_attn.k_proj.safetensors b/out_tensor/model.layers.22.self_attn.k_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..40e358bd8f167beaf2ed10b055512a852e193780 --- /dev/null +++ b/out_tensor/model.layers.22.self_attn.k_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a7172b68f44b31d922042c85b582b39a3937423aa8d7e7bc12d31bfa67931fe8 +size 1788556 diff --git a/out_tensor/model.layers.22.self_attn.o_proj.safetensors b/out_tensor/model.layers.22.self_attn.o_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..b474952bd9cfd870eb11536c58de2499034004c5 --- /dev/null +++ b/out_tensor/model.layers.22.self_attn.o_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e6e3b5a07b7e7ba0a84f5289e01dc1e5ded23e9f669c47907a825ef6dbb7685d +size 3566208 diff --git a/out_tensor/model.layers.22.self_attn.q_proj.safetensors b/out_tensor/model.layers.22.self_attn.q_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..1b37fac4e39d31bb8ed902951e67ac6cc8dba419 --- /dev/null +++ b/out_tensor/model.layers.22.self_attn.q_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e6021712b1592e70e45ef7eb0e8a37dbcfe93d9233c8498fdb29d5c663fc6199 +size 3567244 diff --git a/out_tensor/model.layers.22.self_attn.v_proj.safetensors b/out_tensor/model.layers.22.self_attn.v_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..0b3d82300d4ffa01897fa500fb74e365685d18ce --- /dev/null +++ b/out_tensor/model.layers.22.self_attn.v_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4146df4420e0c2cc8064c2e0ea66f43fbf17226e03ac7491baa0428bf7b4c989 +size 1788556 diff --git a/out_tensor/model.layers.23.mlp.down_proj.safetensors b/out_tensor/model.layers.23.mlp.down_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3a9b47af57fd3ce21e6c0b11e1f44a28dc792dbf --- /dev/null +++ b/out_tensor/model.layers.23.mlp.down_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e5a31796ae7c4ddb7b8f1c5ded7cef78a0c2d4f5ee181e7977ade30d202d7a2d +size 11304464 diff --git a/out_tensor/model.layers.23.mlp.gate_proj.safetensors b/out_tensor/model.layers.23.mlp.gate_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..afee98d2f686c20c437963161b3ab5dbe5b142e8 --- /dev/null +++ b/out_tensor/model.layers.23.mlp.gate_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2bdac808a1ea788629c999eec912a5786252fdc77b28aab9fe97d57ffb85dc19 +size 10709636 diff --git a/out_tensor/model.layers.23.mlp.up_proj.safetensors b/out_tensor/model.layers.23.mlp.up_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..cd939e6861d06ccd63a651047216a6334b8ae65f --- /dev/null +++ b/out_tensor/model.layers.23.mlp.up_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c3d7a46c533129f18402cb420527e7fa703a55835c5bd6e189b3bbed10224a40 +size 10958784 diff --git a/out_tensor/model.layers.23.self_attn.k_proj.safetensors b/out_tensor/model.layers.23.self_attn.k_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..04b07e1f350236d58fc1ecac7375cf5bb3c8d7c5 --- /dev/null +++ b/out_tensor/model.layers.23.self_attn.k_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8d50164446bf83824a51399fe51f20c708e781cf8d08fe69bc665ec9cb430c90 +size 1554384 diff --git a/out_tensor/model.layers.23.self_attn.o_proj.safetensors b/out_tensor/model.layers.23.self_attn.o_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..43320f6e6b9bb1b568c04de7b0fb113cce5f09c2 --- /dev/null +++ b/out_tensor/model.layers.23.self_attn.o_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:eca705739fc140dcda294cd6e1d235b4a500696416c495cc1a1385eac5f1e12c +size 3096480 diff --git a/out_tensor/model.layers.23.self_attn.q_proj.safetensors b/out_tensor/model.layers.23.self_attn.q_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..f16654ee52452c43d6b41121f597c5fdf5972185 --- /dev/null +++ b/out_tensor/model.layers.23.self_attn.q_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0e04faa4417ea2dfa8b6399787a5ef1a4e19c4b273231576e3d265954e0bb249 +size 3098576 diff --git a/out_tensor/model.layers.23.self_attn.v_proj.safetensors b/out_tensor/model.layers.23.self_attn.v_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..0a61f7fc77e50b96900f99def15399bcd2b009cb --- /dev/null +++ b/out_tensor/model.layers.23.self_attn.v_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a289c34bbd8910070a1b2b222eccaf2ee8ebf587264a5ef31b52cdad3c7cb52b +size 1816528 diff --git a/out_tensor/model.layers.24.mlp.down_proj.safetensors b/out_tensor/model.layers.24.mlp.down_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..29cca60917696f10d33212d44d2160ecb19f20d0 --- /dev/null +++ b/out_tensor/model.layers.24.mlp.down_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:926d3aae8bd8de8a421fa07d7ed7873b84c3ceb07ef7e0a562ce47073ad0ecb9 +size 11540696 diff --git a/out_tensor/model.layers.24.mlp.gate_proj.safetensors b/out_tensor/model.layers.24.mlp.gate_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..13cc925fb30822ca271fdf22a83d90518b995175 --- /dev/null +++ b/out_tensor/model.layers.24.mlp.gate_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dd4359ad607d84ee402399cb764d91ca3b09074be53c71de13c89eb5e55a791c +size 10958792 diff --git a/out_tensor/model.layers.24.mlp.up_proj.safetensors b/out_tensor/model.layers.24.mlp.up_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..00b7e007a2fd4fcd0946a67baea2c414cf74333f --- /dev/null +++ b/out_tensor/model.layers.24.mlp.up_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1e1580f2c995bdfd0b4378216f078e4107581165c7d1ee8e62675f4784fd97a8 +size 10958784 diff --git a/out_tensor/model.layers.24.self_attn.k_proj.safetensors b/out_tensor/model.layers.24.self_attn.k_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..64ade2fd06edf20e3d390925a596a3c41a803a4b --- /dev/null +++ b/out_tensor/model.layers.24.self_attn.k_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8bbeabebfac7c3552860ca4c457810cea971ade0a5ba31ebcebc25a3665e55f0 +size 1554384 diff --git a/out_tensor/model.layers.24.self_attn.o_proj.safetensors b/out_tensor/model.layers.24.self_attn.o_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..42ba47552d05a8be46cc3398751aba9f16d65d50 --- /dev/null +++ b/out_tensor/model.layers.24.self_attn.o_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ec8b14d831684cc7f40840c72a6483b88c7d3a4578394d4830fdc7af01402139 +size 3096480 diff --git a/out_tensor/model.layers.24.self_attn.q_proj.safetensors b/out_tensor/model.layers.24.self_attn.q_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..1fcc54c33c86a129fdc1ec8de3f70ad05827e533 --- /dev/null +++ b/out_tensor/model.layers.24.self_attn.q_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:606fd390b7ffc9a817660bb89c4645f0c4d13cc51222b6b435488e2ba74ff1cf +size 3098576 diff --git a/out_tensor/model.layers.24.self_attn.v_proj.safetensors b/out_tensor/model.layers.24.self_attn.v_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..d5302564accb21a16aaf55c20d1ebfc6a29b5d9e --- /dev/null +++ b/out_tensor/model.layers.24.self_attn.v_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7e9ba703b487aa724d8c116925ff6755d74b2ad78cea6d3d400334f7f17abc48 +size 1816528 diff --git a/out_tensor/model.layers.25.mlp.down_proj.safetensors b/out_tensor/model.layers.25.mlp.down_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..12fc94149d7d5913e2c9956aebac16311f375b8b --- /dev/null +++ b/out_tensor/model.layers.25.mlp.down_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7dca133ee156316fe05e8a9745c1cbe44ac603068cd0bcf4f140e0905c4302b0 +size 11540696 diff --git a/out_tensor/model.layers.25.mlp.gate_proj.safetensors b/out_tensor/model.layers.25.mlp.gate_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..fa004815468f2ddd375697ea579fa3beb333f0a2 --- /dev/null +++ b/out_tensor/model.layers.25.mlp.gate_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d34725e3d7342516147234e8cad45a4f8252219e997dd865a522feb40c14af67 +size 10958792 diff --git a/out_tensor/model.layers.25.mlp.up_proj.safetensors b/out_tensor/model.layers.25.mlp.up_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..81d48adf5e631eb2f74bdfcc24c08555979197c2 --- /dev/null +++ b/out_tensor/model.layers.25.mlp.up_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d95e4ef0cc88edc57275952db3bb41777b35ac80714a6f6a2274e1c63568a713 +size 10958784 diff --git a/out_tensor/model.layers.25.self_attn.k_proj.safetensors b/out_tensor/model.layers.25.self_attn.k_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..a48d4321da90f6602a6d291327a021a71a03fd6a --- /dev/null +++ b/out_tensor/model.layers.25.self_attn.k_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7fc9e4199a94be56bece95d8b1d4791033d6f20b1e0420c7f2e77be52ab7cf10 +size 1240824 diff --git a/out_tensor/model.layers.25.self_attn.o_proj.safetensors b/out_tensor/model.layers.25.self_attn.o_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..8d71b5b03c514cc4dae0a3d457790e0782fa4720 --- /dev/null +++ b/out_tensor/model.layers.25.self_attn.o_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1766f8adceaac1e5a02a8f543d99947a7db4798236515d791fe6db32b3097d3c +size 2478816 diff --git a/out_tensor/model.layers.25.self_attn.q_proj.safetensors b/out_tensor/model.layers.25.self_attn.q_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..0c8d11edef887346306e253107221cd33d9c1ce5 --- /dev/null +++ b/out_tensor/model.layers.25.self_attn.q_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:55982939c9fd1599c0cf33a48e6071a536d91b82b9f1a9983fc68010da38c732 +size 2471672 diff --git a/out_tensor/model.layers.25.self_attn.v_proj.safetensors b/out_tensor/model.layers.25.self_attn.v_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..b21f4cfddbb9e79c9af71f3961c125d91df8cc04 --- /dev/null +++ b/out_tensor/model.layers.25.self_attn.v_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fe6b96949ca3305f693e57cfc3e51087f86f93f566bbbc06ebf5fb33a7f9bd68 +size 1502968 diff --git a/out_tensor/model.layers.3.mlp.down_proj.safetensors b/out_tensor/model.layers.3.mlp.down_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..85bd75afb0bc626f4e4ecb344e95452ab60ef25e --- /dev/null +++ b/out_tensor/model.layers.3.mlp.down_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2838ba31f8376255d0091e747a138dc2f48300de4c406dd598f4f8489e2ca984 +size 14323928 diff --git a/out_tensor/model.layers.3.mlp.gate_proj.safetensors b/out_tensor/model.layers.3.mlp.gate_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..6eb38471b2ad96df7c04a9257ca4125bb82b65a4 --- /dev/null +++ b/out_tensor/model.layers.3.mlp.gate_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d1ec0812d0a3f1c90c5b9e75e54573c358cf7b77868dca52f6780a574d5eb42a +size 13907904 diff --git a/out_tensor/model.layers.3.mlp.up_proj.safetensors b/out_tensor/model.layers.3.mlp.up_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..2c5ef62a65a97cb6e8e7d72085bb652e3cf6139d --- /dev/null +++ b/out_tensor/model.layers.3.mlp.up_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:483e56fc8b9185772056e206240c624fef5a63d4be892d4d1d9f565f5d146b66 +size 14276536 diff --git a/out_tensor/model.layers.3.self_attn.k_proj.safetensors b/out_tensor/model.layers.3.self_attn.k_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..86001d170f940b7db9e91925aa7a778ec1e4ec0a --- /dev/null +++ b/out_tensor/model.layers.3.self_attn.k_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:167bab19e4955006882ced43a8a0aa88651c2f05d441c96c3c8970a8b7764840 +size 1554376 diff --git a/out_tensor/model.layers.3.self_attn.o_proj.safetensors b/out_tensor/model.layers.3.self_attn.o_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..358f81312bb414348d69ce171b29bd93f8b815e1 --- /dev/null +++ b/out_tensor/model.layers.3.self_attn.o_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:14026ef8a42b979eb74b35c15edb7c89f5327853defb880632d3c1532e808b33 +size 3096472 diff --git a/out_tensor/model.layers.3.self_attn.q_proj.safetensors b/out_tensor/model.layers.3.self_attn.q_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..18a8a606e5c49510acbb8a76a7770767ea00d7e4 --- /dev/null +++ b/out_tensor/model.layers.3.self_attn.q_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:20b0777b2511c5b7d1b81d2c4445c71c563da014d40d62f01ba204669fa3b478 +size 3098568 diff --git a/out_tensor/model.layers.3.self_attn.v_proj.safetensors b/out_tensor/model.layers.3.self_attn.v_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..b23652545b50cf8f3ba4bdbc4c77d3e456927225 --- /dev/null +++ b/out_tensor/model.layers.3.self_attn.v_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c2e08af931ce6a3bf879c84030d965b7606c437271d3ee226c74b0e83703a0cf +size 1816520 diff --git a/out_tensor/model.layers.4.mlp.down_proj.safetensors b/out_tensor/model.layers.4.mlp.down_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..1708216e117ca5fd0a9115c6810fbfb9abac2492 --- /dev/null +++ b/out_tensor/model.layers.4.mlp.down_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7358cc3c609699d9c065ec29e11950f6267dae82d444fa646f6f2b4bdc8792cf +size 14323928 diff --git a/out_tensor/model.layers.4.mlp.gate_proj.safetensors b/out_tensor/model.layers.4.mlp.gate_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..1c4bf60db7b8bfc63e7a6bcc30500d7c1e2083b3 --- /dev/null +++ b/out_tensor/model.layers.4.mlp.gate_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b3113292b4dd44e849dc0457a0739fa7d8d6d36f3b8a8970f650c666378b06ba +size 13907904 diff --git a/out_tensor/model.layers.4.mlp.up_proj.safetensors b/out_tensor/model.layers.4.mlp.up_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..c743dec961ea05370610561139a16eb1c8f52ce2 --- /dev/null +++ b/out_tensor/model.layers.4.mlp.up_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:34ed634688595c024ea256af3317911ba68c67837db38227e42aed8258922217 +size 14276536 diff --git a/out_tensor/model.layers.4.self_attn.k_proj.safetensors b/out_tensor/model.layers.4.self_attn.k_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..ab78aef94534a89339d3624840a4df08691675d7 --- /dev/null +++ b/out_tensor/model.layers.4.self_attn.k_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b3cdb51421faf3158beaed8e4519d996945b7f8649426913b561a793cd8b4dd7 +size 1788548 diff --git a/out_tensor/model.layers.4.self_attn.o_proj.safetensors b/out_tensor/model.layers.4.self_attn.o_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..dfcfb1d21469ed19dc7237787f4f3f38381de118 --- /dev/null +++ b/out_tensor/model.layers.4.self_attn.o_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4df1271d381bec47d25736d366790feb132625cdc156872e0af016544483c3db +size 3566200 diff --git a/out_tensor/model.layers.4.self_attn.q_proj.safetensors b/out_tensor/model.layers.4.self_attn.q_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..27346c4a8fc69907f581749b693d816ae6389459 --- /dev/null +++ b/out_tensor/model.layers.4.self_attn.q_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:83130936c4f47c4643f05ae8feeb6ed140c105caa0622ee9885be046f562980d +size 3567236 diff --git a/out_tensor/model.layers.4.self_attn.v_proj.safetensors b/out_tensor/model.layers.4.self_attn.v_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..58d8a36e7004c8883d1dfe5af81a8a3bc3409bc9 --- /dev/null +++ b/out_tensor/model.layers.4.self_attn.v_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1a95a4bd8acef0039f5ce6ea75f20694c8f6d33384c46f4445f98b73059dfc13 +size 1788548 diff --git a/out_tensor/model.layers.5.mlp.down_proj.safetensors b/out_tensor/model.layers.5.mlp.down_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..cbe08993d33d4079e64fd5ab95e79f730493f508 --- /dev/null +++ b/out_tensor/model.layers.5.mlp.down_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:930bd81afaf21cc358e1d9a03b28380cbb2dd533c8526d4389dde82ebbdd27cf +size 11807960 diff --git a/out_tensor/model.layers.5.mlp.gate_proj.safetensors b/out_tensor/model.layers.5.mlp.gate_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..de965ae04f0dd705fca6e21ea8c41590623f0b37 --- /dev/null +++ b/out_tensor/model.layers.5.mlp.gate_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:004ffe9fdc1c7bb9dc5fec678854680c306ed410242dd3e48a61e52babf7ceae +size 11253696 diff --git a/out_tensor/model.layers.5.mlp.up_proj.safetensors b/out_tensor/model.layers.5.mlp.up_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..9f7b57587a9717d04e7f308e3925601f41d618ab --- /dev/null +++ b/out_tensor/model.layers.5.mlp.up_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bc7cfb28112830a9b200f498d1eead1697c5e8a871a5fc818336e95ac10cd311 +size 11622328 diff --git a/out_tensor/model.layers.5.self_attn.k_proj.safetensors b/out_tensor/model.layers.5.self_attn.k_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..fa884a756859887ef552e30a2d6b521baad9f6d4 --- /dev/null +++ b/out_tensor/model.layers.5.self_attn.k_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:16462203107e5b38586241ab0771bcf7e7b47526affb43a5a55c90c23b7c0633 +size 1526404 diff --git a/out_tensor/model.layers.5.self_attn.o_proj.safetensors b/out_tensor/model.layers.5.self_attn.o_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..f720d18a6e8031f97c3c4a3549c8bb59217ee8eb --- /dev/null +++ b/out_tensor/model.layers.5.self_attn.o_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7399a5469a6fa1fca9640d6038aed258b3fe428da6cdd75562490b64466f2527 +size 3050104 diff --git a/out_tensor/model.layers.5.self_attn.q_proj.safetensors b/out_tensor/model.layers.5.self_attn.q_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..eae94c24b3cf29f4a695d058ddfa08a98952d63f --- /dev/null +++ b/out_tensor/model.layers.5.self_attn.q_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:226132acc2515ac0e329febc76df167730d5eeba403d35c3e4d133b63614c09a +size 3042948 diff --git a/out_tensor/model.layers.5.self_attn.v_proj.safetensors b/out_tensor/model.layers.5.self_attn.v_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..f1e662794206d0a09da16b09ebaa0464d2c71bbd --- /dev/null +++ b/out_tensor/model.layers.5.self_attn.v_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1fd8b8ba5542ee14280ac500ab96c2235a407fe7b0313cd8d49e3494b3d88b7c +size 1788548 diff --git a/out_tensor/model.layers.6.mlp.down_proj.safetensors b/out_tensor/model.layers.6.mlp.down_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..f4b68371ff75382e3bd696656d982239f7dacb49 --- /dev/null +++ b/out_tensor/model.layers.6.mlp.down_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9c57085d5065b07906bc27b1c59eb6eb61f264b25750135a0c5c4667c778ffa4 +size 11807960 diff --git a/out_tensor/model.layers.6.mlp.gate_proj.safetensors b/out_tensor/model.layers.6.mlp.gate_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..ff99b776a155a50db80119c30d80f5c6fc47e6d8 --- /dev/null +++ b/out_tensor/model.layers.6.mlp.gate_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1ff5766972554e28a7af16ad0fb8498a2a1e398b57131e6e81823b712720750a +size 11253696 diff --git a/out_tensor/model.layers.6.mlp.up_proj.safetensors b/out_tensor/model.layers.6.mlp.up_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..19cfc52e32cfdbba0dd01c99c45c97ae9607718d --- /dev/null +++ b/out_tensor/model.layers.6.mlp.up_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b37d67f083c5a81e5a5aa6f906928b73698001d53354cdc67aba88b6174553c1 +size 11622328 diff --git a/out_tensor/model.layers.6.self_attn.k_proj.safetensors b/out_tensor/model.layers.6.self_attn.k_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..10fdf8829f093b679c9eb67a3f957f8c07986bab --- /dev/null +++ b/out_tensor/model.layers.6.self_attn.k_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1cb7bcff2f4c4e307f47695f37ccc60261a8b64ec791939a44542a536fa40844 +size 1554376 diff --git a/out_tensor/model.layers.6.self_attn.o_proj.safetensors b/out_tensor/model.layers.6.self_attn.o_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..22698b443800bbe0b0082c155470f608c5d139b6 --- /dev/null +++ b/out_tensor/model.layers.6.self_attn.o_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0fc03360c4de308d9a8d42484860165ca237157ede2d68608f9734aefc94a453 +size 3096472 diff --git a/out_tensor/model.layers.6.self_attn.q_proj.safetensors b/out_tensor/model.layers.6.self_attn.q_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..72e5049c2bd173619858eae479bf5863ace08175 --- /dev/null +++ b/out_tensor/model.layers.6.self_attn.q_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f2f3194a9053b69c6168833c905e0d9f5d69b7cd3feb276cc9d7e2fbff06f5d8 +size 3098568 diff --git a/out_tensor/model.layers.6.self_attn.v_proj.safetensors b/out_tensor/model.layers.6.self_attn.v_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..6d37bddfe8ee0d6d6c44ef04da1e1d775c72fad3 --- /dev/null +++ b/out_tensor/model.layers.6.self_attn.v_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3b6b62c5418b451d9935b42b7a2c3239fec00a9c994d6fbe6f52a00fd99c3ab7 +size 1816520 diff --git a/out_tensor/model.layers.7.mlp.down_proj.safetensors b/out_tensor/model.layers.7.mlp.down_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..84832b44a5b4eed71049a9929f546f3cd5725768 --- /dev/null +++ b/out_tensor/model.layers.7.mlp.down_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2aecc927073e246132b4a41efa696c4562542e438b4151b9b46999b063619d28 +size 11540696 diff --git a/out_tensor/model.layers.7.mlp.gate_proj.safetensors b/out_tensor/model.layers.7.mlp.gate_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..9cc73fbf69ee1b59dc50c57070e856ddd9457d94 --- /dev/null +++ b/out_tensor/model.layers.7.mlp.gate_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:74f17912470a410a0659fa3a65c7bcd3f2ab4714408400350278107572ddab88 +size 10958784 diff --git a/out_tensor/model.layers.7.mlp.up_proj.safetensors b/out_tensor/model.layers.7.mlp.up_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..fd4e2a8f4f37b9ba138617d5217d05b8f6e64727 --- /dev/null +++ b/out_tensor/model.layers.7.mlp.up_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9d031b27b6c6793138fde472bceeb3a87fa5b3b6c60efcbdd9c6ae8223268e9a +size 10958776 diff --git a/out_tensor/model.layers.7.self_attn.k_proj.safetensors b/out_tensor/model.layers.7.self_attn.k_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..7777cbbbedb99f2a507c8a6cff321f170997b9d1 --- /dev/null +++ b/out_tensor/model.layers.7.self_attn.k_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ff816c2fa9d066faa66a67b1e21071a96e5f0452c2f00ac67d4b47a40ecdfe30 +size 1788548 diff --git a/out_tensor/model.layers.7.self_attn.o_proj.safetensors b/out_tensor/model.layers.7.self_attn.o_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..f8d6fa28f4055b41f9e7dff114b56fe2102d0c09 --- /dev/null +++ b/out_tensor/model.layers.7.self_attn.o_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5c73e92b6f3a49e2f1149ef3157062627eaa1aca8258395bb2d6ab68986a750a +size 3566200 diff --git a/out_tensor/model.layers.7.self_attn.q_proj.safetensors b/out_tensor/model.layers.7.self_attn.q_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..8ebba6690e949463bb2db418daf940ced21bfaae --- /dev/null +++ b/out_tensor/model.layers.7.self_attn.q_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c22aff89a507bfd7ff3d249e573099e1e58ef79475880d901ca72c7f69d18bb7 +size 3567236 diff --git a/out_tensor/model.layers.7.self_attn.v_proj.safetensors b/out_tensor/model.layers.7.self_attn.v_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..4bacce025a99d1fcbe70085ab4e493a57c1fe230 --- /dev/null +++ b/out_tensor/model.layers.7.self_attn.v_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f63e43730b7a87121507dc7ce4c3c59f08d23faec83237b25033e86860f04d37 +size 1788548 diff --git a/out_tensor/model.layers.8.mlp.down_proj.safetensors b/out_tensor/model.layers.8.mlp.down_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..4374856925189e1125fa9f0fc3666c9c7202413b --- /dev/null +++ b/out_tensor/model.layers.8.mlp.down_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f69090276df9a71bd01bae13f02a65a02693fab508c59b7b7c56a401c9098c25 +size 11540696 diff --git a/out_tensor/model.layers.8.mlp.gate_proj.safetensors b/out_tensor/model.layers.8.mlp.gate_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..1a543894480b33f15e2da70fd31e4b3b7a9f7709 --- /dev/null +++ b/out_tensor/model.layers.8.mlp.gate_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4ab599022a8c42667b9b6ab12c1e79c86827bd14d56a231f1fff7494c147d9ce +size 10958784 diff --git a/out_tensor/model.layers.8.mlp.up_proj.safetensors b/out_tensor/model.layers.8.mlp.up_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..59320395ca9dc7f123c4070c248dc0a69c663081 --- /dev/null +++ b/out_tensor/model.layers.8.mlp.up_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c5c750f889703ba0efe1390f9031a898cba43cd1c619e00e68115ba849689f1f +size 10958776 diff --git a/out_tensor/model.layers.8.self_attn.k_proj.safetensors b/out_tensor/model.layers.8.self_attn.k_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..573bee96e5a1c0bf6be3427498045c4b9a5cb785 --- /dev/null +++ b/out_tensor/model.layers.8.self_attn.k_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:00dc5509d0ce18f1461210f287dcc8e39304c49556db8f249e4b96d2268e9e78 +size 1259464 diff --git a/out_tensor/model.layers.8.self_attn.o_proj.safetensors b/out_tensor/model.layers.8.self_attn.o_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..2251ca10d1d435a2beb60166061dcf48f7b1c25f --- /dev/null +++ b/out_tensor/model.layers.8.self_attn.o_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c612e50c8215e97b22af3df0babe3a6fe8df4ca5b39488880961cdcfea1b0518 +size 2506648 diff --git a/out_tensor/model.layers.8.self_attn.q_proj.safetensors b/out_tensor/model.layers.8.self_attn.q_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..e85fd2555f74abcccec0cf1765bad71c7a852cc2 --- /dev/null +++ b/out_tensor/model.layers.8.self_attn.q_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c5e7b29fd8bdb60e50918723c1335289b08e46218e0a8bb292f89b2cfed27fb2 +size 2508744 diff --git a/out_tensor/model.layers.8.self_attn.v_proj.safetensors b/out_tensor/model.layers.8.self_attn.v_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..9fd3bea10273e4a3e7e1a746cd7362f26ce282b1 --- /dev/null +++ b/out_tensor/model.layers.8.self_attn.v_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5a3cfebb6ee1e8a14b112ec7924f696539e86af8be54a26dbb38872487145fa3 +size 1521608 diff --git a/out_tensor/model.layers.9.mlp.down_proj.safetensors b/out_tensor/model.layers.9.mlp.down_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..0f791c04d2d7dd9acbf8881c94031c89dec0e243 --- /dev/null +++ b/out_tensor/model.layers.9.mlp.down_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:83d9df451ca197ce0ed35d3bea6255d0df29523fc9888dde4a141fd45a117686 +size 11540696 diff --git a/out_tensor/model.layers.9.mlp.gate_proj.safetensors b/out_tensor/model.layers.9.mlp.gate_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..e7b1fe5bc228357eb482ba5aa099328cc9a39f82 --- /dev/null +++ b/out_tensor/model.layers.9.mlp.gate_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e89ceb4391d35aa298799a6fae6673db3d8c7169c30c79a04465a2b24a125cdb +size 10958784 diff --git a/out_tensor/model.layers.9.mlp.up_proj.safetensors b/out_tensor/model.layers.9.mlp.up_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..4f8c042d92f8a6b2b90c33988ac012833da47b44 --- /dev/null +++ b/out_tensor/model.layers.9.mlp.up_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:522e7c3df04a2b2566f2775deb51767ef21f114b627ed81c3e13311870778c99 +size 10958776 diff --git a/out_tensor/model.layers.9.self_attn.k_proj.safetensors b/out_tensor/model.layers.9.self_attn.k_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..9b8ef6b89ebbc073690b67949051919d9dea22ca --- /dev/null +++ b/out_tensor/model.layers.9.self_attn.k_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:27fbace9a1deaa29d81b8922bb85d41be67c8cb696b0e8a13997e84b10f5cc7d +size 1526404 diff --git a/out_tensor/model.layers.9.self_attn.o_proj.safetensors b/out_tensor/model.layers.9.self_attn.o_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..0714929f72d52d4923b687e2f9ea85ca912b1f43 --- /dev/null +++ b/out_tensor/model.layers.9.self_attn.o_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:40f3cf9b19a0edff07836fe286343f8ba82e172bc31a5900e62cfbda8737e274 +size 3050104 diff --git a/out_tensor/model.layers.9.self_attn.q_proj.safetensors b/out_tensor/model.layers.9.self_attn.q_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..eecd8366f30a2abf0a34bed081a4ed51c1b74347 --- /dev/null +++ b/out_tensor/model.layers.9.self_attn.q_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:46f0d61b15bdb70959e7396e9c671a7d67dc2a1d5181bc2f240e52b9395c4ef5 +size 3042948 diff --git a/out_tensor/model.layers.9.self_attn.v_proj.safetensors b/out_tensor/model.layers.9.self_attn.v_proj.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..7a0933ef13a5888dbe81c2dc5ae471d5cc601b88 --- /dev/null +++ b/out_tensor/model.layers.9.self_attn.v_proj.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:979eddc4597d6acb0da69a50933e26a382adcc7196c2b54f762bca5ab694a0a7 +size 1788548 diff --git a/output.safetensors b/output.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..642414862c527c20ed423af928cb3ff3a8a57027 --- /dev/null +++ b/output.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:12f28c18474248772372c36f009d2d0033e0230f1225516ee158ae2cfad60f1f +size 2788067488