{ "bench_version": 1, "track": "tabular", "model": "mlp_tiny", "metric_direction": "lower is better", "n_seeds": 8, "baseline": { "best_cfg": { "lr": 0.01 }, "sweep": [ { "cfg": { "lr": 0.001 }, "mean": 18.323792934417725 }, { "cfg": { "lr": 0.003 }, "mean": 12.731719493865967 }, { "cfg": { "lr": 0.01 }, "mean": 7.733822226524353 } ], "full": { "mean": 7.905456006526947, "std": 0.7333165741197974, "per_seed": [ 7.984025478363037, 6.387940406799316, 7.856008529663086, 8.707314491271973, 8.079574584960938, 7.502847671508789, 8.969817161560059, 7.756119728088379 ], "n": 8 }, "all_union_rates": { "0.001": { "mean": 19.335745573043823, "std": 3.657827404086414, "per_seed": [ 25.777250289916992, 13.66278076171875, 17.315614700317383, 16.539525985717773, 18.554529190063477, 18.02096939086914, 22.761363983154297, 22.053930282592773 ], "n": 8 }, "0.003": { "mean": 13.304376482963562, "std": 1.3247165411283601, "per_seed": [ 14.347658157348633, 10.474533081054688, 13.201133728027344, 12.903553009033203, 13.815519332885742, 13.35650634765625, 12.940177917480469, 15.395930290222168 ], "n": 8 }, "0.01": { "mean": 7.905456006526947, "std": 0.7333165741197974, "per_seed": [ 7.984025478363037, 6.387940406799316, 7.856008529663086, 8.707314491271973, 8.079574584960938, 7.502847671508789, 8.969817161560059, 7.756119728088379 ], "n": 8 } } }, "idea": { "mean": 11.219475030899048, "std": 0.9330729672086647, "per_seed": [ 11.737688064575195, 9.196637153625488, 10.628690719604492, 11.08810806274414, 11.466021537780762, 11.766623497009277, 11.28360366821289, 12.588427543640137 ], "n": 8 }, "comparison": { "delta_mean": 3.314019024372101, "idea_wins": 0, "n_pairs": 8, "per_seed_diffs": [ 3.753662586212158, 2.808696746826172, 2.7726821899414062, 2.380793571472168, 3.386446952819824, 4.263775825500488, 2.313786506652832, 4.832307815551758 ], "p_value": 0.0081, "mde": 0.7627905507582391, "mde_rel_pct": 9.648912727215984, "verdict": "idea worse (significant)", "system_worked": false }, "mechanism_signature": { "prediction": "weakly supported coordinates receive smaller update metric", "observed": { "support_fraction": 0.514455258846283, "q_mean": 0.4733452796936035, "grad_norm": 44.26974844932556 }, "confirmed": true }, "idea_all_union_rates": { "0.001": { "mean": 198.77429580688477, "std": 13.133327131145185, "per_seed": [ 210.48277282714844, 190.67105102539062, 208.48660278320312, 183.82296752929688, 205.18133544921875, 176.4065399169922, 198.54098510742188, 216.60211181640625 ], "n": 8 }, "0.003": { "mean": 39.026103258132935, "std": 9.54746173172683, "per_seed": [ 48.3967170715332, 26.32292938232422, 37.33447265625, 32.64558410644531, 41.14009094238281, 27.771005630493164, 56.16447067260742, 42.433555603027344 ], "n": 8 }, "0.01": { "mean": 11.219475030899048, "std": 0.9330729672086647, "per_seed": [ 11.737688064575195, 9.196637153625488, 10.628690719604492, 11.08810806274414, 11.466021537780762, 11.766623497009277, 11.28360366821289, 12.588427543640137 ], "n": 8 } }, "idea_best_lr": 0.01 }