{ "bench_version": 1, "track": "dynamics", "model": "rnn_small", "metric_direction": "lower is better", "n_seeds": 8, "baseline": { "best_cfg": { "lr": 0.006, "penalty": 0.0 }, "sweep": [ { "cfg": { "lr": 0.001, "penalty": 0.0 }, "mean": 0.0012014944513794035 }, { "cfg": { "lr": 0.001, "penalty": 0.03 }, "mean": 0.0013431779225356877 }, { "cfg": { "lr": 0.001, "penalty": 0.1 }, "mean": 0.001977995882043615 }, { "cfg": { "lr": 0.001, "penalty": 0.3 }, "mean": 0.005485758418217301 }, { "cfg": { "lr": 0.001, "penalty": 1.0 }, "mean": 0.025594600941985846 }, { "cfg": { "lr": 0.003, "penalty": 0.0 }, "mean": 0.0006355914229061455 }, { "cfg": { "lr": 0.003, "penalty": 0.03 }, "mean": 0.0007585847488371655 }, { "cfg": { "lr": 0.003, "penalty": 0.1 }, "mean": 0.0014228892396204174 }, { "cfg": { "lr": 0.003, "penalty": 0.3 }, "mean": 0.00530714844353497 }, { "cfg": { "lr": 0.003, "penalty": 1.0 }, "mean": 0.031064636074006557 }, { "cfg": { "lr": 0.006, "penalty": 0.0 }, "mean": 0.0003905312332790345 }, { "cfg": { "lr": 0.006, "penalty": 0.03 }, "mean": 0.0005399620640673675 }, { "cfg": { "lr": 0.006, "penalty": 0.1 }, "mean": 0.0013079957861918956 }, { "cfg": { "lr": 0.006, "penalty": 0.3 }, "mean": 0.005471856915391982 }, { "cfg": { "lr": 0.006, "penalty": 1.0 }, "mean": 0.03726070327684283 } ], "full": { "mean": 0.00038071037488407455, "std": 7.927794626057588e-05, "per_seed": [ 0.0003860906872432679, 0.0004608586023095995, 0.00025328659103251994, 0.00046188905253075063, 0.00042785078403539956, 0.00024962733732536435, 0.00042033527279272676, 0.0003857446718029678 ], "n": 8 } }, "idea": { "mean": 0.0021949997608317062, "std": 0.0004374952005319115, "per_seed": [ 0.0026410226710140705, 0.0018085994524881244, 0.00271115405485034, 0.0028186936397105455, 0.001880061230622232, 0.0017161071300506592, 0.0017505508149042726, 0.002233809093013406 ], "n": 8 }, "comparison": { "delta_mean": 0.0018142893859476317, "idea_wins": 0, "n_pairs": 8, "per_seed_diffs": [ 0.0022549319837708026, 0.0013477408501785249, 0.00245786746381782, 0.002356804587179795, 0.0014522104465868324, 0.0014664797927252948, 0.0013302155421115458, 0.001848064421210438 ], "p_value": 0.0081, "mde": 0.00040074238262426, "mde_rel_pct": 105.26174463889649, "verdict": "idea worse (significant)", "system_worked": false }, "selected_idea_cfg": { "lr": 0.006, "alpha": 0.02 }, "math_check": { "max_projection_error": 0.0, "interior_drift_error": 5.551115123125783e-17, "constant_positive_drift_cap_steps": 63 }, "track_justification": "Dynamics is structurally matched: controlled pendulum rollout and terminal feasibility.", "mechanism_signature": { "formula": "delta_lambda = alpha*(vbar-tau) away from projection", "per_seed": [ { "seed": 0, "mean_violation": 0.0825553372502327, "mean_lambda": 0.20099468181282282, "drift_prediction_mae": 4.7076746711652336e-18, "drift_observed": 0.001723537926399518, "drift_predicted": 0.0017235379263995185, "cap_fraction": 0.0 }, { "seed": 1, "mean_violation": 0.0704090990126133, "mean_lambda": 0.17491563167423013, "drift_prediction_mae": 3.7114563654548425e-18, "drift_observed": 0.001494265170620074, "drift_predicted": 0.001494265170620074, "cap_fraction": 0.0 }, { "seed": 2, "mean_violation": 0.07129735685884953, "mean_lambda": 0.16706302325800054, "drift_prediction_mae": 4.001007183741646e-18, "drift_observed": 0.0014390677878899228, "drift_predicted": 0.0014390677878899233, "cap_fraction": 0.0 }, { "seed": 3, "mean_violation": 0.06800611577928066, "mean_lambda": 0.16455769450590002, "drift_prediction_mae": 3.9042034183411545e-18, "drift_observed": 0.0014052169518693086, "drift_predicted": 0.0014052169518693096, "cap_fraction": 0.0 }, { "seed": 4, "mean_violation": 0.077877695299685, "mean_lambda": 0.1931413715332746, "drift_prediction_mae": 4.34713442491807e-18, "drift_observed": 0.0016449038421232547, "drift_predicted": 0.0016449038421232547, "cap_fraction": 0.0 }, { "seed": 5, "mean_violation": 0.0745898948982358, "mean_lambda": 0.19240893951803442, "drift_prediction_mae": 4.721442317799971e-18, "drift_observed": 0.0016378540495440126, "drift_predicted": 0.001637854049544013, "cap_fraction": 0.0 }, { "seed": 6, "mean_violation": 0.07793739642947912, "mean_lambda": 0.1956254186630249, "drift_prediction_mae": 4.275714758000374e-18, "drift_observed": 0.0016655272277929485, "drift_predicted": 0.0016655272277929481, "cap_fraction": 0.0 }, { "seed": 7, "mean_violation": 0.06042268052697182, "mean_lambda": 0.15612414248287676, "drift_prediction_mae": 3.4690167129962788e-18, "drift_observed": 0.0013308829530363985, "drift_predicted": 0.001330882953036399, "cap_fraction": 0.0 } ], "observed_mean_drift": 0.0015426569886594298, "predicted_mean_drift": 0.00154265698865943, "mean_drift_prediction_mae": 4.142206231552196e-18, "confirmed": true } }