{ "d_sae": 4096, "actual_l0": 160.5682338, "effective_l0": 86, "cross_entropy": 412.2860979980469, "sae_type": "saebench", "sae_release": "adamkarvonen/saebench_gemma-2-2b_width-2pow12_date-0108", "num_tokens": 10000000, "num_epochs": 1, "lr": 0.0005, "loaded_dag": null, "allow_negative_influence": false, "history": { "train_loss": [ 652.5572274344308, 499.44583718436104, 463.9018794468471, 446.57359662737167, 436.2666741507394, 428.9288805280413, 422.9645320347377, 419.42323325020925, 416.3030757359096, 413.64500841413223, 413.609593636649 ], "train_steps": [ 140, 280, 420, 560, 700, 840, 980, 1120, 1260, 1400, 1407 ], "val_loss": [ 530.6949403320313, 476.5447771484375, 453.5318955078125, 440.38813940429685, 431.7596919921875, 425.6127008300781, 421.003733984375, 417.4211789550781, 414.55859731445315, 412.24590415039063, 412.2860979980469 ], "val_cross_entropy": [ 530.6949403320313, 476.5447771484375, 453.5318955078125, 440.38813940429685, 431.7596919921875, 425.6127008300781, 421.003733984375, 417.4211789550781, 414.55859731445315, 412.24590415039063, 412.2860979980469 ], "effective_l0": [ 118, 103, 97, 93, 91, 89, 88, 87, 86, 86, 86 ], "val_steps": [ 140, 280, 420, 560, 700, 840, 980, 1120, 1260, 1400, 1407 ] }, "sae_id": "MatryoshkaBatchTopK_gemma-2-2b__0108/resid_post_layer_12/trainer_3" }