{ "bench_version": 1, "track": "dynamics", "model": "rnn_small", "metric_direction": "lower is better", "n_seeds": 8, "baseline": { "best_cfg": { "lr": 0.01 }, "sweep": [ { "cfg": { "lr": 0.001 }, "mean": 0.02100143244024366 }, { "cfg": { "lr": 0.003 }, "mean": 0.0030831422773189843 }, { "cfg": { "lr": 0.01 }, "mean": 0.002158296701963991 } ], "full": { "mean": 0.0020533509959932417, "std": 0.0007424383658550215, "per_seed": [ 0.002525518648326397, 0.0009679240174591541, 0.0023394303862005472, 0.0028003137558698654, 0.0015804190188646317, 0.003060286631807685, 0.002158003393560648, 0.0009949121158570051 ], "n": 8 } }, "idea": { "mean": 0.07345113484188914, "std": 0.011031271199206346, "per_seed": [ 0.06164592131972313, 0.07005804777145386, 0.08214986324310303, 0.09056025743484497, 0.05593807250261307, 0.07245660573244095, 0.08499487489461899, 0.06980543583631516 ], "n": 8 }, "comparison": { "delta_mean": 0.0713977838458959, "idea_wins": 0, "n_pairs": 8, "per_seed_diffs": [ 0.05912040267139673, 0.0690901237539947, 0.07981043285690248, 0.0877599436789751, 0.054357653483748436, 0.06939631910063326, 0.08283687150105834, 0.06881052372045815 ], "p_value": 0.0081, "mde": 0.009619511434093848, "mde_rel_pct": 468.4786698847229, "verdict": "idea worse (significant)", "system_worked": false }, "mechanism_signature": { "mechanism_signature": { "predicted": "trajectory velocity should vary with segment/time and integration should reconstruct endpoint", "baseline_endpoint_mse": 0.003087950637564063, "idea_endpoint_mse": 0.05978367105126381, "initial_velocity_mse": 0.04817172512412071, "confirmed": true }, "idea_sweep": [ { "cfg": { "lr": 0.01 }, "mean": 0.07610352244228125 }, { "cfg": { "lr": 0.001 }, "mean": 0.10371190309524536 }, { "cfg": { "lr": 0.01 }, "mean": 0.07610352244228125 } ], "track_match": "dynamics matches controlled solver/refinement trajectories" }, "math_check": { "interpolation_identity_max_error": 0.0, "velocity_identity_max_error": 4.678964614868164e-06, "prediction": "piecewise interpolation and segment velocity are exact" }, "baseline_union_grid": [ { "lr": 0.001 }, { "lr": 0.003 }, { "lr": 0.01 } ], "idea_best_cfg": { "lr": 0.01 } }