{ "bench_version": 1, "track": "dynamics", "model": "rnn_small", "metric_direction": "lower is better", "n_seeds": 8, "baseline": { "best_cfg": { "lr": 0.01, "weight_decay": 0.0, "epochs": 20 }, "sweep": [ { "cfg": { "lr": 0.001, "weight_decay": 0.0, "epochs": 20 }, "mean": 0.00010979082981066313 }, { "cfg": { "lr": 0.003, "weight_decay": 0.0, "epochs": 20 }, "mean": 1.23220872865204e-05 }, { "cfg": { "lr": 0.01, "weight_decay": 0.0, "epochs": 20 }, "mean": 5.326571056230023e-06 } ], "full": { "mean": 6.485018303692414e-06, "std": 3.0896387556205093e-06, "per_seed": [ 5.860124929313315e-06, 6.373980340868002e-06, 2.9345453640416963e-06, 6.137633590697078e-06, 7.82904498919379e-06, 4.8856131797947455e-06, 4.085992713953601e-06, 1.3773211321677081e-05 ], "n": 8 } }, "idea": { "mean": 1.4861759154882748e-05, "std": 7.4365681821899514e-06, "per_seed": [ 6.5364615693397354e-06, 1.7686747014522552e-05, 1.725437687127851e-05, 2.9500635719159618e-05, 1.0894022125285119e-05, 1.0834011845872737e-05, 5.783244432677748e-06, 2.0404573660925962e-05 ], "n": 8 }, "comparison": { "delta_mean": 8.376740851190334e-06, "idea_wins": 0, "n_pairs": 8, "per_seed_diffs": [ 6.763366400264204e-07, 1.131276667365455e-05, 1.4319831507236813e-05, 2.336300212846254e-05, 3.064977136091329e-06, 5.948398666077992e-06, 1.6972517187241465e-06, 6.631362339248881e-06 ], "p_value": 0.0081, "mde": 6.396115118271008e-06, "mde_rel_pct": 98.62909892835944, "verdict": "idea worse (significant)", "system_worked": false }, "mechanism_signature": { "prediction": "trained final states have negative cubic radial coefficient and relax toward finite norm", "trained_model_signature": { "norm_mean": 0.7238094657659531, "norm_std": 0.4153854250907898, "radial_coeff_mean": -0.0551277962513268, "radial_coeff_std": 0.006850266479887068, "radial_velocity_mean": 0.6813768893480301, "predicted_radius_scalar": 1.0, "predicted_relaxation_rate": 2.0, "radial_upper_bound": -0.015625, "confirmed": true }, "idea_trials": [ { "cfg": { "lr": 0.001, "epochs": 20, "lam": 1.0, "dt": 0.05, "cubic_scale": 1.0 }, "result": { "mean": 3.4410899388603866e-05, "std": 5.650175225500919e-06, "per_seed": [ 2.6207977498415858e-05, 2.940478952950798e-05, 3.0609477107645944e-05, 4.4194846850587055e-05, 4.1339764720760286e-05, 3.6304791137808934e-05, 3.362560528330505e-05, 3.3599942980799824e-05 ], "n": 8 } }, { "cfg": { "lr": 0.003, "epochs": 20, "lam": 1.0, "dt": 0.05, "cubic_scale": 1.0 }, "result": { "mean": 1.609290268334007e-05, "std": 4.353488098019092e-06, "per_seed": [ 1.1640918273769785e-05, 1.4350282071973197e-05, 1.364862600894412e-05, 2.4214104996644892e-05, 2.2511794668389484e-05, 1.5585039363941178e-05, 1.4041733265912626e-05, 1.2750722817145288e-05 ], "n": 8 } }, { "cfg": { "lr": 0.01, "epochs": 20, "lam": 1.0, "dt": 0.05, "cubic_scale": 1.0 }, "result": { "mean": 1.4861759154882748e-05, "std": 7.4365681821899514e-06, "per_seed": [ 6.5364615693397354e-06, 1.7686747014522552e-05, 1.725437687127851e-05, 2.9500635719159618e-05, 1.0894022125285119e-05, 1.0834011845872737e-05, 5.783244432677748e-06, 2.0404573660925962e-05 ], "n": 8 } } ], "custom_track": null } }