trainer_config.json

{
    "trainer": {
        "trainer_class": "BatchTopKCrossCoderTrainer",
        "dict_class": "BatchTopKCrossCoder",
        "lr": 0.0001,
        "steps": 97656,
        "auxk_alpha": 0.03125,
        "warmup_steps": 1000,
        "decay_start": null,
        "threshold_beta": 0.999,
        "threshold_start_step": 1000,
        "top_k_aux": 1152,
        "seed": null,
        "activation_dim": 2304,
        "dict_size": 73728,
        "k": 100,
        "sparsity_loss_type": "LossType.CROSSCODER",
        "sparsity_loss_alpha_sae": 1.0,
        "sparsity_loss_alpha_cc": 0.1,
        "device": "cuda",
        "layer": 13,
        "lm_name": "google/gemma-2-2b-it-google/gemma-2-2b",
        "wandb_name": "gemma-2-2b-L13-k100-lr1e-04-local-shuffling-CCLoss",
        "submodule_name": null
    }
}