| {"model_name": "pythia-70m-deduped", "model_class_name": "HookedTransformer", "hook_name": "blocks.3.ln2.hook_normalized", "hook_eval": "NOT_IN_USE", "hook_layer": 3, "hook_head_index": null, "dataset_path": "apollo-research/monology-pile-uncopyrighted-tokenizer-EleutherAI-gpt-neox-20b", "dataset_trust_remote_code": true, "streaming": true, "is_dataset_tokenized": true, "context_size": 2048, "use_cached_activations": false, "cached_activations_path": null, "architecture": "standard", "d_in": 512, "d_sae": 32768, "b_dec_init_method": "zeros", "expansion_factor": 64, "activation_fn": "topk", "activation_fn_kwargs": {"k": 32}, "normalize_sae_decoder": true, "noise_scale": 0.0, "from_pretrained_path": null, "apply_b_dec_to_input": false, "decoder_orthogonal_init": false, "decoder_heuristic_init": false, "init_encoder_as_decoder_transpose": true, "n_batches_in_buffer": 32, "training_tokens": 300003328, "finetuning_tokens": 0, "store_batch_size_prompts": 16, "train_batch_size_tokens": 4096, "normalize_activations": "none", "seqpos_slice": [null], "device": "cuda", "act_store_device": "cuda", "seed": 42, "dtype": "float32", "prepend_bos": true, "autocast": true, "autocast_lm": true, "compile_llm": false, "llm_compilation_mode": null, "compile_sae": false, "sae_compilation_mode": null, "adam_beta1": 0.9, "adam_beta2": 0.999, "mse_loss_normalization": null, "l1_coefficient": 0, "lp_norm": 1, "scale_sparsity_penalty_by_decoder_norm": false, "l1_warm_up_steps": 0, "use_jacobian_loss": true, "jacobian_coefficient": 0.1, "jacobian_warm_up_steps": 3662, "mlp_out_mse_coefficient": 1.0, "lr": 0.0005, "lr_scheduler_name": "constant", "lr_warm_up_steps": 732, "lr_end": 5e-05, "lr_decay_steps": 14648, "n_restart_cycles": 1, "finetuning_method": null, "use_ghost_grads": false, "feature_sampling_window": 1000, "dead_feature_window": 1000, "dead_feature_threshold": 1e-06, "n_eval_batches": 10, "eval_batch_size_prompts": 8, "log_to_wandb": true, "log_activations_store_to_wandb": false, "log_optimizer_state_to_wandb": false, "wandb_project": "jacobian_saes_test", "wandb_id": null, "run_name": "Layer3-32768-J0.1-LR5.0e-04-k32-T3.0e+08", "wandb_entity": null, "wandb_log_frequency": 30, "eval_every_n_wandb_logs": 20, "resume": false, "n_checkpoints": 0, "checkpoint_path": "checkpoints/7l3l5rha", "verbose": true, "model_kwargs": {}, "model_from_pretrained_kwargs": {"center_writing_weights": false}, "tokens_per_buffer": 268435456} |