{ "bench_version": 1, "track": "dynamics", "model": "rnn_small_kinetic_vs_raw", "metric_direction": "lower is better", "n_seeds": 8, "baseline": { "best_cfg": { "lr": 0.006, "h": null }, "sweep": [ { "cfg": { "lr": 0.0015, "h": null }, "mean": 0.003373457526322454 }, { "cfg": { "lr": 0.003, "h": null }, "mean": 0.0019945994718000293 }, { "cfg": { "lr": 0.006, "h": null }, "mean": 0.0011413919273763895 } ], "full": { "mean": 0.001020140029140748, "std": 0.00017823609091726823, "per_seed": [ 0.0009858898120000958, 0.00120925169903785, 0.0009726743446663022, 0.00139775185380131, 0.000904685293789953, 0.0009852343937382102, 0.0008830014267005026, 0.0008226314093917608 ], "n": 8 }, "union_lr_grid": [ 0.0015, 0.003, 0.006 ] }, "idea": { "mean": 0.0013235262595117092, "std": 0.0005778315605493169, "per_seed": [ 0.0010875554289668798, 0.0014675897546112537, 0.001237712800502777, 0.002764510689303279, 0.00098368467297405, 0.0007464992813766003, 0.001115466933697462, 0.0011851905146613717 ], "n": 8 }, "comparison": { "delta_mean": 0.00030338623037096113, "idea_wins": 1, "n_pairs": 8, "per_seed_diffs": [ 0.000101665616966784, 0.00025833805557340384, 0.0002650384558364749, 0.0013667588355019689, 7.899937918409705e-05, -0.00023873511236160994, 0.00023246550699695945, 0.0003625591052696109 ], "p_value": 0.04965, "mde": 0.00039062517861312155, "mde_rel_pct": 38.29132937192363, "verdict": "idea worse (significant)", "system_worked": false }, "mechanism_signature": { "signature": { "prediction": "x Fourier attenuation approximately exp(-h*k^2) and transport displacement proportional to h*x", "predicted_fourier_attenuation": 0.4867522559599717, "observed_fourier_attenuation": 0.4702589511871338, "attenuation_relative_error": 0.033884392382457294, "predicted_transport_displacement": 1.28, "observed_transport_displacement": -0.13549232482910156, "confirmed": true }, "idea_sweep": [ { "cfg": { "lr": 0.0015, "h": 0.08 }, "mean": 0.008491503656841815 }, { "cfg": { "lr": 0.003, "h": 0.08 }, "mean": 0.002999909804202616 }, { "cfg": { "lr": 0.006, "h": 0.08 }, "mean": 0.0016393421683460474 } ], "best_idea_cfg": { "lr": 0.006, "h": 0.08 }, "parameter_parity": { "baseline": 3585, "idea": 3609 } }, "custom_track": null }