{ "bench_version": 1, "track": "tabular", "model": "mlp_tiny", "metric_direction": "lower is better", "n_seeds": 8, "baseline": { "best_cfg": { "lr": 0.006, "weight_decay": 0.0001 }, "sweep": [ { "cfg": { "lr": 0.0015, "weight_decay": 0.0 }, "mean": 16.81662940979004 }, { "cfg": { "lr": 0.0015, "weight_decay": 0.0001 }, "mean": 16.790626525878906 }, { "cfg": { "lr": 0.003, "weight_decay": 0.0 }, "mean": 13.860755920410156 }, { "cfg": { "lr": 0.003, "weight_decay": 0.0001 }, "mean": 13.839483976364136 }, { "cfg": { "lr": 0.006, "weight_decay": 0.0 }, "mean": 10.442761182785034 }, { "cfg": { "lr": 0.006, "weight_decay": 0.0001 }, "mean": 10.238675355911255 } ], "full": { "mean": 10.678602695465088, "std": 1.2286671921700971, "per_seed": [ 10.056567192077637, 11.58691120147705, 9.562552452087402, 9.74867057800293, 8.789999961853027, 11.738417625427246, 11.37514591217041, 12.570556640625 ], "n": 8 } }, "idea": { "mean": 10.288059711456299, "std": 0.9213815344241648, "per_seed": [ 10.669325828552246, 10.632473945617676, 10.393888473510742, 9.149672508239746, 8.439580917358398, 11.03823184967041, 10.63460636138916, 11.346697807312012 ], "n": 8 }, "comparison": { "delta_mean": -0.39054298400878906, "idea_wins": 6, "n_pairs": 8, "per_seed_diffs": [ 0.6127586364746094, -0.954437255859375, 0.8313360214233398, -0.5989980697631836, -0.3504190444946289, -0.7001857757568359, -0.74053955078125, -1.2238588333129883 ], "p_value": 0.1779, "mde": 0.6140738235288303, "mde_rel_pct": 5.750507262430606, "verdict": "no significant win", "system_worked": false }, "mechanism_signature": { "quantity": "active hidden fraction after trained gate", "predicted": "adaptive shrinkage should reduce active fraction", "observed_mean": 0.999560534954071, "observed_per_seed": [ 0.999804675579071, 0.9994531273841858, 0.9996093511581421, 0.9993749856948853 ], "confirmed": false }, "math_check": { "analytic_divergence": 3.467686653137207, "finite_difference_divergence": 3.466308116912842, "abs_error": 0.0013785362243652344, "confirmed": true }, "idea_sweep": [ { "cfg": { "lr": 0.006, "sure_weight": 0.01 }, "mean": 10.211340188980103 }, { "cfg": { "lr": 0.006, "sure_weight": 0.02 }, "mean": 10.212185859680176 }, { "cfg": { "lr": 0.006, "sure_weight": 0.04 }, "mean": 10.213889598846436 } ] }