{ "bench_version": 1, "track": "dynamics", "model": "rnn_small", "metric_direction": "lower is better", "n_seeds": 8, "baseline": { "best_cfg": { "lr": 0.006 }, "sweep": [ { "cfg": { "lr": 0.001 }, "mean": 0.0002686133266252 }, { "cfg": { "lr": 0.003 }, "mean": 2.2429956516134553e-05 }, { "cfg": { "lr": 0.006 }, "mean": 1.2719097412627889e-05 } ], "full": { "mean": 1.3684802866009704e-05, "std": 2.4297331089709535e-06, "per_seed": [ 1.0965275578200817e-05, 1.3063688129477669e-05, 1.363205592497252e-05, 1.321537001786055e-05, 1.694267484708689e-05, 1.790331407391932e-05, 1.3341849808057304e-05, 1.0414194548502564e-05 ], "n": 8 } }, "idea": { "mean": 1.3684802866009704e-05, "std": 2.4297331089709535e-06, "per_seed": [ 1.0965275578200817e-05, 1.3063688129477669e-05, 1.363205592497252e-05, 1.321537001786055e-05, 1.694267484708689e-05, 1.790331407391932e-05, 1.3341849808057304e-05, 1.0414194548502564e-05 ], "n": 8 }, "comparison": { "delta_mean": 0.0, "idea_wins": 0, "n_pairs": 8, "per_seed_diffs": [ 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0 ], "p_value": 1.0, "mde": 0.0, "mde_rel_pct": 0.0, "verdict": "no measurable effect", "system_worked": false }, "mechanism_signature": { "prediction": "contraction-aware perturbation should remain below Euler stability cap alpha<2", "predicted_alpha": 1.8, "observed_mean_lambda_h": 0.2277172260841973, "observed_idea_proxy": 7.496510079363361e-09, "rows": [ { "seed": 0, "estimated_lambda": 3.643026351928711, "mean_step_h": 0.04968743026256561, "baseline_proxy": 6.96672941558063e-09, "idea_proxy": 6.96672941558063e-09 }, { "seed": 1, "estimated_lambda": 4.969603538513184, "mean_step_h": 0.049560677260160446, "baseline_proxy": 7.60337570682168e-09, "idea_proxy": 7.60337570682168e-09 }, { "seed": 2, "estimated_lambda": 4.052803993225098, "mean_step_h": 0.04964262992143631, "baseline_proxy": 1.2150849215686321e-08, "idea_proxy": 1.2150849215686321e-08 }, { "seed": 3, "estimated_lambda": 4.340335845947266, "mean_step_h": 0.049574412405490875, "baseline_proxy": 7.668859325349331e-09, "idea_proxy": 7.668859325349331e-09 }, { "seed": 4, "estimated_lambda": 4.611820220947266, "mean_step_h": 0.049705807119607925, "baseline_proxy": 9.313225746154785e-09, "idea_proxy": 9.313225746154785e-09 }, { "seed": 5, "estimated_lambda": 4.991096496582031, "mean_step_h": 0.049615517258644104, "baseline_proxy": 5.6461431086063385e-09, "idea_proxy": 5.6461431086063385e-09 }, { "seed": 6, "estimated_lambda": 5.737295150756836, "mean_step_h": 0.04948228970170021, "baseline_proxy": 4.016328603029251e-09, "idea_proxy": 4.016328603029251e-09 }, { "seed": 7, "estimated_lambda": 4.377944469451904, "mean_step_h": 0.04963567852973938, "baseline_proxy": 6.606569513678551e-09, "idea_proxy": 6.606569513678551e-09 } ], "confirmed": true }, "math_sanity": { "predicted_boundary_lambda_h": 2.0, "observed_grid_boundary": 1.9999999999999998, "max_boundary_error": 2.220446049250313e-16, "passed": true }, "idea_sweep": [ { "cfg": { "lr": 0.006, "alpha": 1.8, "penalty": 0.1 }, "mean": 1.3684802866009704e-05 }, { "cfg": { "lr": 0.003, "alpha": 0.5, "penalty": 0.1 }, "mean": 2.628253741931985e-05 }, { "cfg": { "lr": 0.001, "alpha": 1.0, "penalty": 0.1 }, "mean": 0.00021658051718986826 } ], "protocol_notes": { "structural_match": "controlled damped pendulum rollout tests stability/control dynamics", "baseline_grid": [ { "lr": 0.001 }, { "lr": 0.003 }, { "lr": 0.006 } ], "idea_grid": [ { "lr": 0.006, "alpha": 1.8, "penalty": 0.1 }, { "lr": 0.003, "alpha": 0.5, "penalty": 0.1 }, { "lr": 0.001, "alpha": 1.0, "penalty": 0.1 } ], "epochs": 15, "n_train": 4000, "n_test": 1000 } }