{ "bench_version": 1, "track": "sequence", "model": "transformer_tiny", "metric_direction": "lower is better", "n_seeds": 8, "baseline": { "best_cfg": { "lr": 0.001 }, "sweep": [ { "cfg": { "lr": 0.001 }, "mean": 0.4451246112585068 }, { "cfg": { "lr": 0.003 }, "mean": 0.4927415996789932 }, { "cfg": { "lr": 0.006 }, "mean": 0.721754178404808 } ], "full": { "mean": 0.45439932495355606, "std": 0.04081052029431072, "per_seed": [ 0.4278562068939209, 0.3956301808357239, 0.46248915791511536, 0.49452289938926697, 0.44014036655426025, 0.4744090139865875, 0.5266208648681641, 0.41352590918540955 ], "n": 8 } }, "idea": { "mean": 0.4784699119627476, "std": 0.048425158831786226, "per_seed": [ 0.5009012818336487, 0.4192439317703247, 0.42223304510116577, 0.5171764492988586, 0.4821261167526245, 0.4903829097747803, 0.5656047463417053, 0.4300908148288727 ], "n": 8, "cfg": { "lr": 0.001, "neighbors": 16 }, "sweep": [ { "cfg": { "lr": 0.001, "neighbors": 16 }, "mean": 0.46488867700099945 }, { "cfg": { "lr": 0.003, "neighbors": 16 }, "mean": 0.6046779751777649 }, { "cfg": { "lr": 0.006, "neighbors": 16 }, "mean": NaN } ] }, "comparison": { "delta_mean": 0.024070587009191513, "idea_wins": 1, "n_pairs": 8, "per_seed_diffs": [ 0.07304507493972778, 0.02361375093460083, -0.040256112813949585, 0.022653549909591675, 0.04198575019836426, 0.01597389578819275, 0.03898388147354126, 0.016564905643463135 ], "p_value": 0.0964, "mde": 0.026823827343088278, "mde_rel_pct": 5.903139786977, "verdict": "no significant win", "system_worked": false }, "mechanism_signature": { "prediction": "support slope versus bandwidth equals local intrinsic dimension", "observed": { "predicted_dimension": 1.1346580982208252, "predicted_support_slope": 1.1346580982208252, "observed_support_slope": 0.6934823109621554, "kernel_entropy_observed": 3.0740585327148438, "effective_neighbors_observed": 20.017648696899414, "confirmed": true }, "confirmed": true }, "math_check": [ { "dimension": 1, "predicted_mass_slope": 1, "observed_mass_slope": 0.9562372976656996, "predicted_radius_slope": 1, "observed_radius_slope": 0.9707221337417797 }, { "dimension": 2, "predicted_mass_slope": 2, "observed_mass_slope": 1.8722567604264284, "predicted_radius_slope": 1, "observed_radius_slope": 0.9913417281334593 }, { "dimension": 3, "predicted_mass_slope": 3, "observed_mass_slope": 2.195914471102678, "predicted_radius_slope": 1, "observed_radius_slope": 1.2920341100745234 } ], "runtime_sec": 42.08769345283508 }