Upload SAE model.layers.12.mlp

model.layers.12.mlp/cfg.json

@@ -1 +1 @@
-{"architecture": "standard", "d_in": 896, "d_sae": 14336, "dtype": "float32", "device": "cuda", "model_name": "bew/qwen2.5-lmsys-fft", "hook_name": "model.layers.12.mlp", "hook_layer": 12, "hook_head_index": null, "activation_fn_str": "relu", "activation_fn_kwargs": {}, "apply_b_dec_to_input": false, "finetuning_scaling_factor": false, "sae_lens_training_version": "5.3.3", "prepend_bos": true, "dataset_path": "bew/lmsys-chat-1m-qwen2.5-instruct-1024-context", "dataset_trust_remote_code": true, "context_size": 256, "normalize_activations": "none", "neuronpedia_id": null, "model_from_pretrained_kwargs": {}, "seqpos_slice": [null], "l1_coefficient": 5, "lp_norm": 1.0, "use_ghost_grads": false, "normalize_sae_decoder": false, "noise_scale": 0.0, "decoder_orthogonal_init": false, "init_encoder_as_decoder_transpose": true, "mse_loss_normalization": null, "decoder_heuristic_init": true, "scale_sparsity_penalty_by_decoder_norm": true, "jumprelu_init_threshold": 0.001, "jumprelu_bandwidth": 0.001}
+{"architecture": "standard", "d_in": 896, "d_sae": 14336, "dtype": "float32", "device": "cuda", "model_name": "bew/qwen2.5-lmsys-fft", "hook_name": "model.layers.12.mlp", "hook_layer": 12, "hook_head_index": null, "activation_fn_str": "relu", "activation_fn_kwargs": {}, "apply_b_dec_to_input": false, "finetuning_scaling_factor": false, "sae_lens_training_version": "5.3.3", "prepend_bos": true, "dataset_path": "bew/lmsys-chat-1m-qwen2.5-instruct-1024-context", "dataset_trust_remote_code": true, "context_size": 512, "normalize_activations": "none", "neuronpedia_id": null, "model_from_pretrained_kwargs": {}, "seqpos_slice": [null], "l1_coefficient": 5, "lp_norm": 1.0, "use_ghost_grads": false, "normalize_sae_decoder": false, "noise_scale": 0.0, "decoder_orthogonal_init": false, "init_encoder_as_decoder_transpose": true, "mse_loss_normalization": null, "decoder_heuristic_init": true, "scale_sparsity_penalty_by_decoder_norm": true, "jumprelu_init_threshold": 0.001, "jumprelu_bandwidth": 0.001}

model.layers.12.mlp/sae_weights.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6918f2bd83dd6017f553bee9e2639bc9adfaa4a1fb9609220ab9bc261706e142
+oid sha256:c3e96c503c7ce8d11170bc60a2445f33f13505b4e9d089f7edf47bc4862d552e
 size 102821696