{ "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": 0.001, "weight_decay": 0.0 }, "mean": 25.60164165496826 }, { "cfg": { "lr": 0.001, "weight_decay": 0.0001 }, "mean": 25.508389472961426 }, { "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": { "cfg": { "lr": 0.006, "reg": 0.02, "alpha": 0.5 }, "mean": 98.17565250396729, "std": 15.026422279731827, "per_seed": [ 108.2091293334961, 109.38567352294922, 88.7021255493164, 95.20142364501953, 67.35282135009766, 112.1347427368164, 89.88146209716797, 114.537841796875 ], "n": 8 }, "comparison": { "delta_mean": 87.4970498085022, "idea_wins": 0, "n_pairs": 8, "per_seed_diffs": [ 98.15256214141846, 97.79876232147217, 79.139573097229, 85.4527530670166, 58.56282138824463, 100.39632511138916, 78.50631618499756, 101.96728515625 ], "p_value": 0.0081, "mde": 12.591954120737624, "mde_rel_pct": 117.91761974705814, "verdict": "idea worse (significant)", "system_worked": false }, "mechanism_signature": { "predicted_vs_observed": [ { "seed": 0, "rho_pred": 0.6043921917676925, "observed_block_response": [ 37.716896057128906, 54.9484977722168, 154.05938720703125 ] }, { "seed": 1, "rho_pred": 0.5927385747432707, "observed_block_response": [ 39.42379379272461, 57.7188835144043, 155.12850952148438 ] }, { "seed": 2, "rho_pred": 0.5786523669958115, "observed_block_response": [ 42.31975173950195, 50.240013122558594, 138.11692810058594 ] }, { "seed": 3, "rho_pred": 0.560906821489334, "observed_block_response": [ 43.896156311035156, 62.861045837402344, 148.82882690429688 ] }, { "seed": 4, "rho_pred": 0.5791306406259537, "observed_block_response": [ 45.193443298339844, 60.726924896240234, 165.0826416015625 ] }, { "seed": 5, "rho_pred": 0.5629717469215392, "observed_block_response": [ 45.11399841308594, 62.12171173095703, 140.21766662597656 ] }, { "seed": 6, "rho_pred": 0.5911920607089995, "observed_block_response": [ 41.594322204589844, 59.584537506103516, 163.77099609375 ] }, { "seed": 7, "rho_pred": 0.5964572012424468, "observed_block_response": [ 46.23796081542969, 56.21791076660156, 158.2962188720703 ] } ], "prediction": "gated blocks should avoid unstable coupling", "confirmed": false, "note": "NN probe is heuristic, not exact Hessian spectral radius" }, "idea_sweep": [ { "cfg": { "lr": 0.001, "reg": 0.02, "alpha": 0.5 }, "mean": 227.52948379516602, "std": 6.764206456146394, "per_seed": [ 234.45175170898438, 228.4759063720703, 232.97796630859375, 214.48780822753906, 232.79397583007812, 225.4188232421875, 219.51397705078125, 232.11566162109375 ], "n": 8 }, { "cfg": { "lr": 0.003, "reg": 0.02, "alpha": 0.5 }, "mean": 198.1618824005127, "std": 7.105863534863813, "per_seed": [ 205.67877197265625, 202.6162872314453, 200.92047119140625, 187.7106170654297, 191.00547790527344, 201.2646484375, 189.39215087890625, 206.70663452148438 ], "n": 8 }, { "cfg": { "lr": 0.006, "reg": 0.02, "alpha": 0.5 }, "mean": 98.17565250396729, "std": 15.026422279731827, "per_seed": [ 108.2091293334961, 109.38567352294922, 88.7021255493164, 95.20142364501953, 67.35282135009766, 112.1347427368164, 89.88146209716797, 114.537841796875 ], "n": 8 } ], "custom_track": null }