{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.992,
  "eval_steps": 100,
  "global_step": 372,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "clip_ratio": 0.0,
      "completion_length": 603.5946578979492,
      "epoch": 0.010666666666666666,
      "grad_norm": 0.5516607761383057,
      "kl": 0.0,
      "learning_rate": 7.894736842105262e-08,
      "loss": 0.02,
      "reward": 0.6678571403026581,
      "reward_std": 0.3216256983578205,
      "rewards/accuracy_reward": 0.6571428701281548,
      "rewards/format_reward": 0.010714285774156451,
      "step": 1
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 597.3767852783203,
      "epoch": 0.05333333333333334,
      "grad_norm": 0.2556845247745514,
      "kl": 0.0001468658447265625,
      "learning_rate": 3.9473684210526315e-07,
      "loss": 0.0409,
      "reward": 0.6642857193946838,
      "reward_std": 0.33807775331661105,
      "rewards/accuracy_reward": 0.6566964322701097,
      "rewards/format_reward": 0.007589285756694153,
      "step": 5
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 589.1785781860351,
      "epoch": 0.10666666666666667,
      "grad_norm": 0.3433384597301483,
      "kl": 0.0002094268798828125,
      "learning_rate": 7.894736842105263e-07,
      "loss": 0.0595,
      "reward": 0.6535714335739613,
      "reward_std": 0.32146761380136013,
      "rewards/accuracy_reward": 0.6446428611874581,
      "rewards/format_reward": 0.008928571571595966,
      "step": 10
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 600.2507247924805,
      "epoch": 0.16,
      "grad_norm": 0.48812174797058105,
      "kl": 0.00038931369781494143,
      "learning_rate": 1.1842105263157894e-06,
      "loss": 0.0511,
      "reward": 0.6946428589522838,
      "reward_std": 0.3123627858236432,
      "rewards/accuracy_reward": 0.6871428646147251,
      "rewards/format_reward": 0.00750000006519258,
      "step": 15
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 596.2403602600098,
      "epoch": 0.21333333333333335,
      "grad_norm": 0.3659345805644989,
      "kl": 0.001643228530883789,
      "learning_rate": 1.5789473684210526e-06,
      "loss": 0.0613,
      "reward": 0.7185714311897755,
      "reward_std": 0.28948054276406765,
      "rewards/accuracy_reward": 0.7135714307427407,
      "rewards/format_reward": 0.005000000027939678,
      "step": 20
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 587.3889312744141,
      "epoch": 0.26666666666666666,
      "grad_norm": 0.460025429725647,
      "kl": 0.0052853584289550785,
      "learning_rate": 1.973684210526316e-06,
      "loss": 0.0622,
      "reward": 0.7282142877578736,
      "reward_std": 0.2650401175022125,
      "rewards/accuracy_reward": 0.7235714301466942,
      "rewards/format_reward": 0.004642857192084193,
      "step": 25
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 614.826789855957,
      "epoch": 0.32,
      "grad_norm": 0.20424464344978333,
      "kl": 0.00449981689453125,
      "learning_rate": 2.368421052631579e-06,
      "loss": 0.0648,
      "reward": 0.7517857104539871,
      "reward_std": 0.2205220595933497,
      "rewards/accuracy_reward": 0.7489285707473755,
      "rewards/format_reward": 0.002857142873108387,
      "step": 30
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 576.6128578186035,
      "epoch": 0.37333333333333335,
      "grad_norm": 0.1937570720911026,
      "kl": 0.006718826293945312,
      "learning_rate": 2.763157894736842e-06,
      "loss": 0.0462,
      "reward": 0.79035714417696,
      "reward_std": 0.1855543740093708,
      "rewards/accuracy_reward": 0.7892857149243355,
      "rewards/format_reward": 0.0010714285774156452,
      "step": 35
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 598.0246505737305,
      "epoch": 0.4266666666666667,
      "grad_norm": 0.23902535438537598,
      "kl": 0.01360149383544922,
      "learning_rate": 2.9997345912364375e-06,
      "loss": 0.0454,
      "reward": 0.7460714317858219,
      "reward_std": 0.21041248124092818,
      "rewards/accuracy_reward": 0.7392857171595096,
      "rewards/format_reward": 0.006785714346915483,
      "step": 40
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 582.4003646850585,
      "epoch": 0.48,
      "grad_norm": 0.6725406050682068,
      "kl": 0.017615890502929686,
      "learning_rate": 2.996749821181634e-06,
      "loss": 0.0359,
      "reward": 0.7892857164144516,
      "reward_std": 0.2517069520428777,
      "rewards/accuracy_reward": 0.75571428835392,
      "rewards/format_reward": 0.03357142908498645,
      "step": 45
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 603.898217010498,
      "epoch": 0.5333333333333333,
      "grad_norm": 0.5048671364784241,
      "kl": 0.011526107788085938,
      "learning_rate": 2.9904551426434754e-06,
      "loss": 0.0335,
      "reward": 0.7896428570151329,
      "reward_std": 0.2881605923175812,
      "rewards/accuracy_reward": 0.7271428614854812,
      "rewards/format_reward": 0.06250000121071934,
      "step": 50
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 580.8153602600098,
      "epoch": 0.5866666666666667,
      "grad_norm": 0.6043830513954163,
      "kl": 0.01994476318359375,
      "learning_rate": 2.980864475656959e-06,
      "loss": 0.0254,
      "reward": 0.8578571483492852,
      "reward_std": 0.32197347320616243,
      "rewards/accuracy_reward": 0.74178571626544,
      "rewards/format_reward": 0.11607143122237176,
      "step": 55
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 579.2271438598633,
      "epoch": 0.64,
      "grad_norm": 0.29581740498542786,
      "kl": 0.012204742431640625,
      "learning_rate": 2.9679990289969723e-06,
      "loss": 0.0351,
      "reward": 0.8657142907381058,
      "reward_std": 0.2968884438276291,
      "rewards/accuracy_reward": 0.761785714328289,
      "rewards/format_reward": 0.10392857452388853,
      "step": 60
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 591.5560768127441,
      "epoch": 0.6933333333333334,
      "grad_norm": 0.4318093955516815,
      "kl": 0.03280029296875,
      "learning_rate": 2.951887253277264e-06,
      "loss": 0.0297,
      "reward": 0.9300000056624412,
      "reward_std": 0.3733633913099766,
      "rewards/accuracy_reward": 0.7417857185006141,
      "rewards/format_reward": 0.18821429125964642,
      "step": 65
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 611.58857421875,
      "epoch": 0.7466666666666667,
      "grad_norm": 2.673933267593384,
      "kl": 0.048126220703125,
      "learning_rate": 2.9325647780348364e-06,
      "loss": 0.0394,
      "reward": 0.9132142946124077,
      "reward_std": 0.3908999715000391,
      "rewards/accuracy_reward": 0.7196428544819355,
      "rewards/format_reward": 0.19357143295928836,
      "step": 70
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 605.3753646850586,
      "epoch": 0.8,
      "grad_norm": 0.6781793236732483,
      "kl": 0.059979248046875,
      "learning_rate": 2.9100743329388826e-06,
      "loss": 0.042,
      "reward": 0.8521428614854812,
      "reward_std": 0.44031637236475946,
      "rewards/accuracy_reward": 0.6639285758137703,
      "rewards/format_reward": 0.18821429088711739,
      "step": 75
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 611.6189346313477,
      "epoch": 0.8533333333333334,
      "grad_norm": 2.177452564239502,
      "kl": 0.1802734375,
      "learning_rate": 2.884465653298514e-06,
      "loss": 0.0659,
      "reward": 0.6303571477532387,
      "reward_std": 0.5197178602218628,
      "rewards/accuracy_reward": 0.4832142934203148,
      "rewards/format_reward": 0.14714286094531417,
      "step": 80
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 681.2253623962403,
      "epoch": 0.9066666666666666,
      "grad_norm": 5.2435994148254395,
      "kl": 0.60625,
      "learning_rate": 2.8557953700782305e-06,
      "loss": 0.1428,
      "reward": 0.4467857230454683,
      "reward_std": 0.4360184334218502,
      "rewards/accuracy_reward": 0.41107143461704254,
      "rewards/format_reward": 0.03571428642608225,
      "step": 85
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 697.349642944336,
      "epoch": 0.96,
      "grad_norm": 17.68018913269043,
      "kl": 2.22822265625,
      "learning_rate": 2.8241268846643613e-06,
      "loss": 0.2796,
      "reward": 0.49678572118282316,
      "reward_std": 0.43236474245786666,
      "rewards/accuracy_reward": 0.46214286610484123,
      "rewards/format_reward": 0.034642857825383545,
      "step": 90
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 650.735001373291,
      "epoch": 1.0213333333333334,
      "grad_norm": 16.169004440307617,
      "kl": 4.5029296875,
      "learning_rate": 2.789530228659411e-06,
      "loss": 0.3745,
      "reward": 0.49964286386966705,
      "reward_std": 0.42348831966519357,
      "rewards/accuracy_reward": 0.4578571505844593,
      "rewards/format_reward": 0.041785715334117415,
      "step": 95
    },
    {
      "epoch": 1.0746666666666667,
      "grad_norm": 16.28914451599121,
      "learning_rate": 2.7520819090143655e-06,
      "loss": 0.2175,
      "step": 100
    },
    {
      "epoch": 1.0746666666666667,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 548.8102988891602,
      "eval_kl": 12.394375,
      "eval_loss": 0.5909453630447388,
      "eval_reward": 0.5816285777390003,
      "eval_reward_std": 0.37753888714313505,
      "eval_rewards/accuracy_reward": 0.5438000068366527,
      "eval_rewards/format_reward": 0.03782857222296297,
      "eval_runtime": 6124.2952,
      "eval_samples_per_second": 0.816,
      "eval_steps_per_second": 0.012,
      "step": 100
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 550.720006942749,
      "epoch": 1.1280000000000001,
      "grad_norm": 49.118804931640625,
      "kl": 4.5041015625,
      "learning_rate": 2.711864738841427e-06,
      "loss": 0.3096,
      "reward": 0.6191071469336749,
      "reward_std": 0.38678194154053924,
      "rewards/accuracy_reward": 0.5803571455180645,
      "rewards/format_reward": 0.038750000752042976,
      "step": 105
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 523.8392898559571,
      "epoch": 1.1813333333333333,
      "grad_norm": 19.96983528137207,
      "kl": 4.0896484375,
      "learning_rate": 2.668967654281324e-06,
      "loss": 0.2402,
      "reward": 0.6832142889499664,
      "reward_std": 0.36785929203033446,
      "rewards/accuracy_reward": 0.6496428608894348,
      "rewards/format_reward": 0.03357142929453403,
      "step": 110
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 554.866431427002,
      "epoch": 1.2346666666666666,
      "grad_norm": 6.781939506530762,
      "kl": 3.09013671875,
      "learning_rate": 2.6234855178301717e-06,
      "loss": 0.1646,
      "reward": 0.6760714322328567,
      "reward_std": 0.310880384221673,
      "rewards/accuracy_reward": 0.6560714334249497,
      "rewards/format_reward": 0.020000000251457095,
      "step": 115
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 536.1167854309082,
      "epoch": 1.288,
      "grad_norm": 6.253053665161133,
      "kl": 2.498974609375,
      "learning_rate": 2.5755189085608046e-06,
      "loss": 0.1272,
      "reward": 0.6971428595483303,
      "reward_std": 0.3085736159235239,
      "rewards/accuracy_reward": 0.6796428576111794,
      "rewards/format_reward": 0.017500000167638065,
      "step": 120
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 526.4314331054687,
      "epoch": 1.3413333333333333,
      "grad_norm": 3.0490851402282715,
      "kl": 2.29580078125,
      "learning_rate": 2.5251738997024913e-06,
      "loss": 0.1268,
      "reward": 0.7439285695552826,
      "reward_std": 0.31444666273891925,
      "rewards/accuracy_reward": 0.7060714334249496,
      "rewards/format_reward": 0.037857143627479675,
      "step": 125
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 546.8860710144043,
      "epoch": 1.3946666666666667,
      "grad_norm": 10.747838973999023,
      "kl": 1.85048828125,
      "learning_rate": 2.4725618240708804e-06,
      "loss": 0.108,
      "reward": 0.7599999979138374,
      "reward_std": 0.3282610267400742,
      "rewards/accuracy_reward": 0.7164285719394684,
      "rewards/format_reward": 0.043571429373696444,
      "step": 130
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 534.8628593444824,
      "epoch": 1.448,
      "grad_norm": 3.6572587490081787,
      "kl": 1.6306396484375,
      "learning_rate": 2.417799027866917e-06,
      "loss": 0.0736,
      "reward": 0.7350000008940697,
      "reward_std": 0.35150096751749516,
      "rewards/accuracy_reward": 0.6803571470081806,
      "rewards/format_reward": 0.05464285863563419,
      "step": 135
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 530.9917877197265,
      "epoch": 1.5013333333333332,
      "grad_norm": 8.864882469177246,
      "kl": 2.42578125,
      "learning_rate": 2.3610066133891706e-06,
      "loss": 0.0744,
      "reward": 0.7300000011920929,
      "reward_std": 0.4424170255661011,
      "rewards/accuracy_reward": 0.6117857187986374,
      "rewards/format_reward": 0.1182142885401845,
      "step": 140
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 516.1107185363769,
      "epoch": 1.5546666666666666,
      "grad_norm": 3.990010976791382,
      "kl": 2.644091796875,
      "learning_rate": 2.3023101712285398e-06,
      "loss": 0.057,
      "reward": 0.8574999988079071,
      "reward_std": 0.49391784705221653,
      "rewards/accuracy_reward": 0.6457142919301987,
      "rewards/format_reward": 0.21178571931086482,
      "step": 145
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 524.2689338684082,
      "epoch": 1.608,
      "grad_norm": 2.412954092025757,
      "kl": 2.77197265625,
      "learning_rate": 2.241839502537563e-06,
      "loss": 0.0402,
      "reward": 0.8878571495413781,
      "reward_std": 0.5054053217172623,
      "rewards/accuracy_reward": 0.6367857187986374,
      "rewards/format_reward": 0.2510714347474277,
      "step": 150
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 517.0146469116211,
      "epoch": 1.6613333333333333,
      "grad_norm": 3.1102499961853027,
      "kl": 2.2763671875,
      "learning_rate": 2.179728331988501e-06,
      "loss": 0.0304,
      "reward": 0.8621428608894348,
      "reward_std": 0.44099100194871427,
      "rewards/accuracy_reward": 0.6635714322328568,
      "rewards/format_reward": 0.19857143410481512,
      "step": 155
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 499.7432174682617,
      "epoch": 1.7146666666666666,
      "grad_norm": 6.359482288360596,
      "kl": 2.4619140625,
      "learning_rate": 2.116114012054961e-06,
      "loss": 0.0077,
      "reward": 0.8435714364051818,
      "reward_std": 0.48649194166064264,
      "rewards/accuracy_reward": 0.6239285781979561,
      "rewards/format_reward": 0.21964286239817737,
      "step": 160
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 498.7282180786133,
      "epoch": 1.768,
      "grad_norm": 9.238311767578125,
      "kl": 3.5078125,
      "learning_rate": 2.0511372192710126e-06,
      "loss": 0.0425,
      "reward": 0.9250000014901161,
      "reward_std": 0.5103498436510563,
      "rewards/accuracy_reward": 0.6107142917811871,
      "rewards/format_reward": 0.3142857207916677,
      "step": 165
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 491.2950042724609,
      "epoch": 1.8213333333333335,
      "grad_norm": 9.016805648803711,
      "kl": 2.2509765625,
      "learning_rate": 1.984941643139478e-06,
      "loss": -0.0307,
      "reward": 1.0628571465611458,
      "reward_std": 0.5711491458117962,
      "rewards/accuracy_reward": 0.6128571502864361,
      "rewards/format_reward": 0.45000001005828383,
      "step": 170
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 489.2357177734375,
      "epoch": 1.8746666666666667,
      "grad_norm": 18.03326988220215,
      "kl": 3.234765625,
      "learning_rate": 1.9176736683773613e-06,
      "loss": -0.0073,
      "reward": 1.2264285877346992,
      "reward_std": 0.5304918490350247,
      "rewards/accuracy_reward": 0.6496428653597832,
      "rewards/format_reward": 0.5767857208848,
      "step": 175
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 480.22750396728514,
      "epoch": 1.928,
      "grad_norm": 5.03889799118042,
      "kl": 2.7779296875,
      "learning_rate": 1.8494820512010797e-06,
      "loss": -0.0156,
      "reward": 1.2835714370012283,
      "reward_std": 0.5365196898579597,
      "rewards/accuracy_reward": 0.6614285789430141,
      "rewards/format_reward": 0.6221428655087948,
      "step": 180
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 504.85750579833984,
      "epoch": 1.9813333333333332,
      "grad_norm": 20.948030471801758,
      "kl": 2.94013671875,
      "learning_rate": 1.780517590367375e-06,
      "loss": -0.0176,
      "reward": 1.2474999949336052,
      "reward_std": 0.5874028220772743,
      "rewards/accuracy_reward": 0.6375000067055225,
      "rewards/format_reward": 0.6100000090897083,
      "step": 185
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 494.6467887878418,
      "epoch": 2.042666666666667,
      "grad_norm": 2.8199145793914795,
      "kl": 2.37197265625,
      "learning_rate": 1.7109327936973479e-06,
      "loss": -0.0157,
      "reward": 1.353214304149151,
      "reward_std": 0.5247410386800766,
      "rewards/accuracy_reward": 0.6764285713434219,
      "rewards/format_reward": 0.6767857156693935,
      "step": 190
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 483.0278602600098,
      "epoch": 2.096,
      "grad_norm": 13.793828010559082,
      "kl": 2.445703125,
      "learning_rate": 1.6408815408210818e-06,
      "loss": -0.0227,
      "reward": 1.3939285814762115,
      "reward_std": 0.5199761673808098,
      "rewards/accuracy_reward": 0.7057142898440361,
      "rewards/format_reward": 0.6882142893970012,
      "step": 195
    },
    {
      "epoch": 2.1493333333333333,
      "grad_norm": 3.724238872528076,
      "learning_rate": 1.5705187428886465e-06,
      "loss": -0.0155,
      "step": 200
    },
    {
      "epoch": 2.1493333333333333,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 512.0718896484375,
      "eval_kl": 1.576041015625,
      "eval_loss": -0.009688925929367542,
      "eval_reward": 1.365571434855461,
      "eval_reward_std": 0.49168447187542913,
      "eval_rewards/accuracy_reward": 0.66702857619524,
      "eval_rewards/format_reward": 0.6985428606867791,
      "eval_runtime": 6003.0557,
      "eval_samples_per_second": 0.833,
      "eval_steps_per_second": 0.012,
      "step": 200
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 505.77339363098145,
      "epoch": 2.2026666666666666,
      "grad_norm": 2.5565314292907715,
      "kl": 1.9068603515625,
      "learning_rate": 1.5e-06,
      "loss": -0.0106,
      "reward": 1.388928584754467,
      "reward_std": 0.5045508090406656,
      "rewards/accuracy_reward": 0.7166071452200413,
      "rewards/format_reward": 0.6723214339464902,
      "step": 205
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 523.7457168579101,
      "epoch": 2.2560000000000002,
      "grad_norm": 3.069352626800537,
      "kl": 1.618798828125,
      "learning_rate": 1.429481257111354e-06,
      "loss": -0.009,
      "reward": 1.4150000035762786,
      "reward_std": 0.49943920969963074,
      "rewards/accuracy_reward": 0.7103571489453315,
      "rewards/format_reward": 0.7046428620815277,
      "step": 210
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 530.0814331054687,
      "epoch": 2.3093333333333335,
      "grad_norm": 3.0543601512908936,
      "kl": 1.9750244140625,
      "learning_rate": 1.3591184591789185e-06,
      "loss": 0.019,
      "reward": 1.4221428513526917,
      "reward_std": 0.4604501351714134,
      "rewards/accuracy_reward": 0.6807142853736877,
      "rewards/format_reward": 0.7414285719394684,
      "step": 215
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 534.6321449279785,
      "epoch": 2.3626666666666667,
      "grad_norm": 2.2766215801239014,
      "kl": 1.20283203125,
      "learning_rate": 1.289067206302653e-06,
      "loss": 0.0046,
      "reward": 1.4853571504354477,
      "reward_std": 0.4559413559734821,
      "rewards/accuracy_reward": 0.7289285689592362,
      "rewards/format_reward": 0.7564285725355149,
      "step": 220
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 535.5114295959472,
      "epoch": 2.416,
      "grad_norm": 6.880006790161133,
      "kl": 1.637890625,
      "learning_rate": 1.2194824096326252e-06,
      "loss": 0.0129,
      "reward": 1.4717857182025909,
      "reward_std": 0.45263413041830064,
      "rewards/accuracy_reward": 0.7039285734295845,
      "rewards/format_reward": 0.7678571432828903,
      "step": 225
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 519.4785758972168,
      "epoch": 2.469333333333333,
      "grad_norm": 1.4145762920379639,
      "kl": 1.39964599609375,
      "learning_rate": 1.1505179487989203e-06,
      "loss": 0.0005,
      "reward": 1.5700000017881393,
      "reward_std": 0.4442869186401367,
      "rewards/accuracy_reward": 0.7607142880558968,
      "rewards/format_reward": 0.809285718202591,
      "step": 230
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 531.5350059509277,
      "epoch": 2.522666666666667,
      "grad_norm": 2.4906344413757324,
      "kl": 2.0687744140625,
      "learning_rate": 1.0823263316226388e-06,
      "loss": 0.0247,
      "reward": 1.4989285796880722,
      "reward_std": 0.45227440148591996,
      "rewards/accuracy_reward": 0.7221428595483304,
      "rewards/format_reward": 0.7767857149243355,
      "step": 235
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 535.0050033569336,
      "epoch": 2.576,
      "grad_norm": 1.7402544021606445,
      "kl": 1.248944091796875,
      "learning_rate": 1.0150583568605221e-06,
      "loss": 0.0079,
      "reward": 1.5596428662538528,
      "reward_std": 0.41812130361795424,
      "rewards/accuracy_reward": 0.7350000031292439,
      "rewards/format_reward": 0.8246428564190864,
      "step": 240
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 537.6096481323242,
      "epoch": 2.6293333333333333,
      "grad_norm": 4.602447509765625,
      "kl": 1.56826171875,
      "learning_rate": 9.488627807289882e-07,
      "loss": 0.0175,
      "reward": 1.525357148051262,
      "reward_std": 0.431475493311882,
      "rewards/accuracy_reward": 0.7296428553760052,
      "rewards/format_reward": 0.7957142800092697,
      "step": 245
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 518.1739295959472,
      "epoch": 2.6826666666666665,
      "grad_norm": 2.083872079849243,
      "kl": 1.552978515625,
      "learning_rate": 8.838859879450389e-07,
      "loss": 0.0217,
      "reward": 1.5639285832643508,
      "reward_std": 0.426826455257833,
      "rewards/accuracy_reward": 0.7457142874598504,
      "rewards/format_reward": 0.8182142853736878,
      "step": 250
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 521.3553634643555,
      "epoch": 2.7359999999999998,
      "grad_norm": 4.696753978729248,
      "kl": 1.580224609375,
      "learning_rate": 8.202716680115e-07,
      "loss": 0.0233,
      "reward": 1.5221428662538528,
      "reward_std": 0.4356484226882458,
      "rewards/accuracy_reward": 0.7357142850756645,
      "rewards/format_reward": 0.7864285722374916,
      "step": 255
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 513.2957183837891,
      "epoch": 2.7893333333333334,
      "grad_norm": 3.655956745147705,
      "kl": 1.584326171875,
      "learning_rate": 7.581604974624371e-07,
      "loss": 0.0256,
      "reward": 1.4928571552038192,
      "reward_std": 0.4848907835781574,
      "rewards/accuracy_reward": 0.7232142873108387,
      "rewards/format_reward": 0.7696428582072258,
      "step": 260
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 509.24785842895506,
      "epoch": 2.8426666666666667,
      "grad_norm": 2.87070894241333,
      "kl": 1.654150390625,
      "learning_rate": 6.976898287714604e-07,
      "loss": 0.0159,
      "reward": 1.4539285823702812,
      "reward_std": 0.4541813228279352,
      "rewards/accuracy_reward": 0.6950000040233135,
      "rewards/format_reward": 0.7589285731315613,
      "step": 265
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 525.0035736083985,
      "epoch": 2.896,
      "grad_norm": 3.930969476699829,
      "kl": 1.767578125,
      "learning_rate": 6.389933866108296e-07,
      "loss": 0.0307,
      "reward": 1.4810714304447175,
      "reward_std": 0.4880278453230858,
      "rewards/accuracy_reward": 0.724285714328289,
      "rewards/format_reward": 0.7567857146263123,
      "step": 270
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 525.0428611755372,
      "epoch": 2.9493333333333336,
      "grad_norm": 2.4287071228027344,
      "kl": 1.80546875,
      "learning_rate": 5.822009721330832e-07,
      "loss": 0.0282,
      "reward": 1.4832143098115922,
      "reward_std": 0.4480956181883812,
      "rewards/accuracy_reward": 0.7142857179045677,
      "rewards/format_reward": 0.7689285710453987,
      "step": 275
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 511.9317855834961,
      "epoch": 3.010666666666667,
      "grad_norm": 2.061641216278076,
      "kl": 2.241748046875,
      "learning_rate": 5.2743817592912e-07,
      "loss": 0.0328,
      "reward": 1.455000001192093,
      "reward_std": 0.46853253729641436,
      "rewards/accuracy_reward": 0.6910714313387871,
      "rewards/format_reward": 0.763928571343422,
      "step": 280
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 522.1992881774902,
      "epoch": 3.064,
      "grad_norm": 1.9162739515304565,
      "kl": 1.256103515625,
      "learning_rate": 4.7482610029750927e-07,
      "loss": 0.0036,
      "reward": 1.5257142961025238,
      "reward_std": 0.42034803740680216,
      "rewards/accuracy_reward": 0.7310714319348335,
      "rewards/format_reward": 0.7946428567171097,
      "step": 285
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 530.8821426391602,
      "epoch": 3.1173333333333333,
      "grad_norm": 2.2770814895629883,
      "kl": 2.02861328125,
      "learning_rate": 4.244810914391956e-07,
      "loss": -0.001,
      "reward": 1.5200000017881394,
      "reward_std": 0.4468880720436573,
      "rewards/accuracy_reward": 0.7317857146263123,
      "rewards/format_reward": 0.7882142812013626,
      "step": 290
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 530.2903610229492,
      "epoch": 3.1706666666666665,
      "grad_norm": 2.8380885124206543,
      "kl": 1.5606689453125,
      "learning_rate": 3.7651448216982855e-07,
      "loss": 0.0219,
      "reward": 1.534642866253853,
      "reward_std": 0.42025899738073347,
      "rewards/accuracy_reward": 0.7257142871618271,
      "rewards/format_reward": 0.8089285686612129,
      "step": 295
    },
    {
      "epoch": 3.224,
      "grad_norm": 2.1940526962280273,
      "learning_rate": 3.3103234571867633e-07,
      "loss": 0.0001,
      "step": 300
    },
    {
      "epoch": 3.224,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 519.2250689086914,
      "eval_kl": 1.42582421875,
      "eval_loss": 0.013943231664597988,
      "eval_reward": 1.5034000079631806,
      "eval_reward_std": 0.41947181399166583,
      "eval_rewards/accuracy_reward": 0.6849714323282242,
      "eval_rewards/format_reward": 0.8184285705089569,
      "eval_runtime": 6122.2221,
      "eval_samples_per_second": 0.817,
      "eval_steps_per_second": 0.012,
      "step": 300
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 524.1357151031494,
      "epoch": 3.2773333333333334,
      "grad_norm": 2.4619476795196533,
      "kl": 1.54144287109375,
      "learning_rate": 2.8813526115857293e-07,
      "loss": 0.0063,
      "reward": 1.5721428662538528,
      "reward_std": 0.40110522620379924,
      "rewards/accuracy_reward": 0.7501785714179278,
      "rewards/format_reward": 0.8219642840325833,
      "step": 305
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 509.5435729980469,
      "epoch": 3.3306666666666667,
      "grad_norm": 8.40831470489502,
      "kl": 1.7404296875,
      "learning_rate": 2.479180909856347e-07,
      "loss": 0.0136,
      "reward": 1.4957142978906632,
      "reward_std": 0.4558893218636513,
      "rewards/accuracy_reward": 0.7010714299976826,
      "rewards/format_reward": 0.7946428552269935,
      "step": 310
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 520.3792892456055,
      "epoch": 3.384,
      "grad_norm": 4.6018877029418945,
      "kl": 1.7560302734375,
      "learning_rate": 2.104697713405892e-07,
      "loss": 0.0214,
      "reward": 1.5135714292526246,
      "reward_std": 0.4333773460239172,
      "rewards/accuracy_reward": 0.7192857176065445,
      "rewards/format_reward": 0.79428571164608,
      "step": 315
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 494.4396492004395,
      "epoch": 3.437333333333333,
      "grad_norm": 4.831773281097412,
      "kl": 1.40341796875,
      "learning_rate": 1.7587311533563887e-07,
      "loss": 0.0066,
      "reward": 1.5846428632736207,
      "reward_std": 0.4081498969346285,
      "rewards/accuracy_reward": 0.7592857137322426,
      "rewards/format_reward": 0.825357137620449,
      "step": 320
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 508.63000259399416,
      "epoch": 3.490666666666667,
      "grad_norm": 2.3977479934692383,
      "kl": 1.6579345703125,
      "learning_rate": 1.4420462992176975e-07,
      "loss": 0.0181,
      "reward": 1.5185714453458785,
      "reward_std": 0.44710296392440796,
      "rewards/accuracy_reward": 0.7267857134342194,
      "rewards/format_reward": 0.7917857110500336,
      "step": 325
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 519.3960723876953,
      "epoch": 3.544,
      "grad_norm": 2.906923532485962,
      "kl": 1.84501953125,
      "learning_rate": 1.1553434670148605e-07,
      "loss": 0.0087,
      "reward": 1.503214305639267,
      "reward_std": 0.46401076950132847,
      "rewards/accuracy_reward": 0.7150000005960464,
      "rewards/format_reward": 0.7882142841815949,
      "step": 330
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 515.9103630065918,
      "epoch": 3.5973333333333333,
      "grad_norm": 2.045380115509033,
      "kl": 1.571337890625,
      "learning_rate": 8.992566706111727e-08,
      "loss": -0.0024,
      "reward": 1.540000006556511,
      "reward_std": 0.435180502384901,
      "rewards/accuracy_reward": 0.7353571504354477,
      "rewards/format_reward": 0.8046428561210632,
      "step": 335
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 509.41107177734375,
      "epoch": 3.6506666666666665,
      "grad_norm": 3.562167167663574,
      "kl": 1.6155029296875,
      "learning_rate": 6.743522196516388e-08,
      "loss": 0.0079,
      "reward": 1.53071428835392,
      "reward_std": 0.42104902639985087,
      "rewards/accuracy_reward": 0.7242857120931149,
      "rewards/format_reward": 0.8064285680651665,
      "step": 340
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 507.3475028991699,
      "epoch": 3.7039999999999997,
      "grad_norm": 8.227246284484863,
      "kl": 1.647607421875,
      "learning_rate": 4.811274672273652e-08,
      "loss": 0.0177,
      "reward": 1.5507142812013626,
      "reward_std": 0.4271882243454456,
      "rewards/accuracy_reward": 0.7385714307427407,
      "rewards/format_reward": 0.8121428564190865,
      "step": 345
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 508.85178833007814,
      "epoch": 3.7573333333333334,
      "grad_norm": 4.183529376983643,
      "kl": 1.48662109375,
      "learning_rate": 3.200097100302812e-08,
      "loss": 0.0025,
      "reward": 1.5217857152223586,
      "reward_std": 0.43769130408763884,
      "rewards/accuracy_reward": 0.7239285737276078,
      "rewards/format_reward": 0.797857141494751,
      "step": 350
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 517.6982177734375,
      "epoch": 3.8106666666666666,
      "grad_norm": 2.569713592529297,
      "kl": 1.5388671875,
      "learning_rate": 1.9135524343040946e-08,
      "loss": 0.0151,
      "reward": 1.522142869234085,
      "reward_std": 0.44272076338529587,
      "rewards/accuracy_reward": 0.7325000062584877,
      "rewards/format_reward": 0.7896428555250168,
      "step": 355
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 508.73714141845704,
      "epoch": 3.864,
      "grad_norm": 3.259504795074463,
      "kl": 1.57216796875,
      "learning_rate": 9.54485735652455e-09,
      "loss": 0.0061,
      "reward": 1.5335714370012283,
      "reward_std": 0.4290111746639013,
      "rewards/accuracy_reward": 0.7246428586542606,
      "rewards/format_reward": 0.808928570151329,
      "step": 360
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 525.433218383789,
      "epoch": 3.9173333333333336,
      "grad_norm": 2.6966588497161865,
      "kl": 1.55546875,
      "learning_rate": 3.2501788183657564e-09,
      "loss": 0.0102,
      "reward": 1.5185714438557625,
      "reward_std": 0.45268934071063993,
      "rewards/accuracy_reward": 0.7200000017881394,
      "rewards/format_reward": 0.7985714256763459,
      "step": 365
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 525.9932167053223,
      "epoch": 3.970666666666667,
      "grad_norm": 2.979252815246582,
      "kl": 1.54794921875,
      "learning_rate": 2.6540876356256906e-10,
      "loss": 0.017,
      "reward": 1.5514285653829574,
      "reward_std": 0.447709359228611,
      "rewards/accuracy_reward": 0.7378571465611458,
      "rewards/format_reward": 0.81357142329216,
      "step": 370
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 454.072322845459,
      "epoch": 3.992,
      "kl": 1.3245849609375,
      "reward": 1.5482143089175224,
      "reward_std": 0.4065241804346442,
      "rewards/accuracy_reward": 0.7303571458905935,
      "rewards/format_reward": 0.8178571350872517,
      "step": 372,
      "total_flos": 0.0,
      "train_loss": 0.046423821178983436,
      "train_runtime": 64088.6589,
      "train_samples_per_second": 0.468,
      "train_steps_per_second": 0.006
    }
  ],
  "logging_steps": 5,
  "max_steps": 372,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": false,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 14,
  "trial_name": null,
  "trial_params": null
}