{ "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": 0.0015202847716864198 }, { "cfg": { "lr": 0.003 }, "mean": 0.0009046953928191215 }, { "cfg": { "lr": 0.01 }, "mean": 0.00043617923438432626 } ], "full": { "mean": 0.0003015457559740753, "std": 0.00019741488644742065, "per_seed": [ 0.0006247494602575898, 0.0004462092474568635, 0.00017418812785763294, 0.0004995701019652188, 0.0003730663738679141, 0.00010612085316097364, 6.512444815598428e-05, 0.00012333743507042527 ], "n": 8 } }, "idea": { "mean": 0.0002918587179010501, "std": 0.00018751789044513985, "per_seed": [ 0.0005810419097542763, 0.0004426969971973449, 0.00017546206072438508, 0.0005002013058401644, 0.00034365191822871566, 0.00010483426012797281, 6.512441905215383e-05, 0.00012185687228338793 ], "n": 8, "sweep": [ { "cfg": { "lr": 0.001, "floor": 1e-05 }, "mean": 0.0015202843060251325, "sweep": { "mean": 0.0015202843060251325, "std": 0.0002737330879148898, "per_seed": [ 0.001321663847193122, 0.001269076718017459, 0.0015261193038895726, 0.0019642773550003767 ], "n": 4 } }, { "cfg": { "lr": 0.003, "floor": 1e-05 }, "mean": 0.0009028831991599873, "sweep": { "mean": 0.0009028831991599873, "std": 0.00015086129898563282, "per_seed": [ 0.0009533172706142068, 0.0007026336970739067, 0.0008414627518504858, 0.0011141190771013498 ], "n": 4 } }, { "cfg": { "lr": 0.01, "floor": 1e-05 }, "mean": 0.00042485056837904267, "sweep": { "mean": 0.00042485056837904267, "std": 0.0001521402244052453, "per_seed": [ 0.0005810419097542763, 0.0004426969971973449, 0.00017546206072438508, 0.0005002013058401644 ], "n": 4 } } ] }, "comparison": { "delta_mean": -9.687038073025178e-06, "idea_wins": 6, "n_pairs": 8, "per_seed_diffs": [ -4.370755050331354e-05, -3.5122502595186234e-06, 1.2739328667521477e-06, 6.312038749456406e-07, -2.9414455639198422e-05, -1.2865930330008268e-06, -2.9103830456733704e-11, -1.4805627870373428e-06 ], "p_value": 0.0823, "mde": 1.4284899018237163e-05, "mde_rel_pct": 4.737224363212485, "verdict": "no significant win", "system_worked": false }, "mechanism_signature": { "track_choice": "dynamics matches stability/control structure", "idea_sweep": [ { "cfg": { "lr": 0.001, "floor": 1e-05 }, "mean": 0.0015202843060251325, "sweep": { "mean": 0.0015202843060251325, "std": 0.0002737330879148898, "per_seed": [ 0.001321663847193122, 0.001269076718017459, 0.0015261193038895726, 0.0019642773550003767 ], "n": 4 } }, { "cfg": { "lr": 0.003, "floor": 1e-05 }, "mean": 0.0009028831991599873, "sweep": { "mean": 0.0009028831991599873, "std": 0.00015086129898563282, "per_seed": [ 0.0009533172706142068, 0.0007026336970739067, 0.0008414627518504858, 0.0011141190771013498 ], "n": 4 } }, { "cfg": { "lr": 0.01, "floor": 1e-05 }, "mean": 0.00042485056837904267, "sweep": { "mean": 0.00042485056837904267, "std": 0.0001521402244052453, "per_seed": [ 0.0005810419097542763, 0.0004426969971973449, 0.00017546206072438508, 0.0005002013058401644 ], "n": 4 } } ], "mechanism_signature": { "prediction": "hard convex interpolation eliminates out-of-domain reconstructed states", "observed": [ { "seed": 0, "raw_invalid_rate": 0.0, "limited_invalid_rate": 0.0, "activation_rate": 0.0, "raw_anchor_displacement": 0.03173816576600075, "limited_anchor_displacement": 0.03173816576600075 }, { "seed": 1, "raw_invalid_rate": 0.0, "limited_invalid_rate": 0.0, "activation_rate": 0.0, "raw_anchor_displacement": 0.03535204008221626, "limited_anchor_displacement": 0.03535204008221626 }, { "seed": 2, "raw_invalid_rate": 0.0, "limited_invalid_rate": 0.0, "activation_rate": 0.0, "raw_anchor_displacement": 0.03403977304697037, "limited_anchor_displacement": 0.03403977304697037 }, { "seed": 3, "raw_invalid_rate": 0.0, "limited_invalid_rate": 0.0, "activation_rate": 0.0, "raw_anchor_displacement": 0.03182002902030945, "limited_anchor_displacement": 0.03182002902030945 }, { "seed": 4, "raw_invalid_rate": 0.0, "limited_invalid_rate": 0.0, "activation_rate": 0.0, "raw_anchor_displacement": 0.03627397492527962, "limited_anchor_displacement": 0.03627397492527962 }, { "seed": 5, "raw_invalid_rate": 0.0, "limited_invalid_rate": 0.0, "activation_rate": 0.0, "raw_anchor_displacement": 0.0337686762213707, "limited_anchor_displacement": 0.0337686762213707 }, { "seed": 6, "raw_invalid_rate": 0.0, "limited_invalid_rate": 0.0, "activation_rate": 0.0, "raw_anchor_displacement": 0.030422382056713104, "limited_anchor_displacement": 0.030422382056713104 }, { "seed": 7, "raw_invalid_rate": 0.0, "limited_invalid_rate": 0.0, "activation_rate": 0.0, "raw_anchor_displacement": 0.0314624160528183, "limited_anchor_displacement": 0.0314624160528183 } ], "raw_invalid_rate_mean": 0.0, "limited_invalid_rate_mean": 0.0, "confirmed": true } } }