{ "bench_version": 1, "track": "dynamics", "model": "rnn_small", "metric_direction": "lower is better", "n_seeds": 8, "baseline": { "best_cfg": { "lr": 0.006, "weight_decay": 0.0001 }, "sweep": [ { "cfg": { "lr": 0.001, "weight_decay": 0.0 }, "mean": 0.005222459672950208 }, { "cfg": { "lr": 0.003, "weight_decay": 0.0 }, "mean": 0.0032230458600679412 }, { "cfg": { "lr": 0.006, "weight_decay": 0.0 }, "mean": 0.0022474244324257597 }, { "cfg": { "lr": 0.001, "weight_decay": 0.0001 }, "mean": 0.005207901762332767 }, { "cfg": { "lr": 0.003, "weight_decay": 0.0001 }, "mean": 0.003198997670551762 }, { "cfg": { "lr": 0.006, "weight_decay": 0.0001 }, "mean": 0.0022117759071988985 } ], "full": { "mean": 0.0022117759071988985, "std": 0.0004126881098400608, "per_seed": [ 0.002800536109134555, 0.001347046927548945, 0.001825663261115551, 0.0023813715670257807, 0.0023169058840721846, 0.0024658460170030594, 0.0022247303277254105, 0.002332107163965702 ], "n": 8 } }, "idea": { "cfg": { "lr": 0.006, "weight_decay": 0.0 }, "mean": 0.0018650132114998996, "std": 0.0008113546405834626, "per_seed": [ 0.0019402806647121906, 0.000917025376111269, 0.00216244300827384, 0.003614878747612238, 0.0023528579622507095, 0.0011243090266361833, 0.0014117673272266984, 0.0013965435791760683 ], "n": 8 }, "comparison": { "delta_mean": -0.00034676269569899887, "idea_wins": 5, "n_pairs": 8, "per_seed_diffs": [ -0.0008602554444223642, -0.00043002155143767595, 0.00033677974715828896, 0.0012335071805864573, 3.595207817852497e-05, -0.0013415369903668761, -0.0008129630004987121, -0.0009355635847896338 ], "p_value": 0.25545, "mde": 0.0007020822623155458, "mde_rel_pct": 31.742920249307595, "verdict": "no significant win", "system_worked": false }, "idea_sweep": [ { "cfg": { "lr": 0.001, "weight_decay": 0.0 }, "mean": 0.005775320241809823, "std": 0.0032488792459768865, "per_seed": [ 0.013064713217318058, 0.001774493488483131, 0.005221209488809109, 0.0029860660433769226, 0.005788332317024469, 0.0035972988698631525, 0.006834390107542276, 0.006936058402061462 ], "n": 8 }, { "cfg": { "lr": 0.003, "weight_decay": 0.0 }, "mean": 0.003850284934742376, "std": 0.0013980431321628994, "per_seed": [ 0.005784670822322369, 0.0026722021866589785, 0.003800347214564681, 0.005372739862650633, 0.004920608829706907, 0.0028360977303236723, 0.0014036931097507477, 0.004011919721961021 ], "n": 8 }, { "cfg": { "lr": 0.006, "weight_decay": 0.0 }, "mean": 0.0018650132114998996, "std": 0.0008113546405834626, "per_seed": [ 0.0019402806647121906, 0.000917025376111269, 0.00216244300827384, 0.003614878747612238, 0.0023528579622507095, 0.0011243090266361833, 0.0014117673272266984, 0.0013965435791760683 ], "n": 8 }, { "cfg": { "lr": 0.001, "weight_decay": 0.0001 }, "mean": 0.005752541357651353, "std": 0.003251318835613531, "per_seed": [ 0.013069597072899342, 0.0017632902599871159, 0.005176736507564783, 0.00296420487575233, 0.005783510394394398, 0.003592958441004157, 0.006851297803223133, 0.006818735506385565 ], "n": 8 }, { "cfg": { "lr": 0.003, "weight_decay": 0.0001 }, "mean": 0.0038360641920007765, "std": 0.0013847205833962984, "per_seed": [ 0.005729560274630785, 0.0026723749469965696, 0.003779245540499687, 0.005330996587872505, 0.004909616895020008, 0.002826679963618517, 0.0013972243759781122, 0.004042814951390028 ], "n": 8 }, { "cfg": { "lr": 0.006, "weight_decay": 0.0001 }, "mean": 0.0018727781061897986, "std": 0.000814390952787726, "per_seed": [ 0.0019529082346707582, 0.0009154137806035578, 0.002177728107199073, 0.0036332407034933567, 0.0023326349910348654, 0.0011026168940588832, 0.0014148338232189417, 0.0014528483152389526 ], "n": 8 } ], "idea_best_signature_metrics": { "test_metric": 0.0019402806647121906, "return_residual_mean": 0.6867897510528564, "floquet_rho": 0.0386572889983654 }, "mechanism_signature": { "baseline_predicted": "rho<1 implies reduced return residual", "predicted_rho": "<1 should be stable", "baseline_rho": 0.03690243512392044, "idea_rho": 0.0386572889983654, "baseline_return_residual": 0.9603143334388733, "idea_return_residual": 0.6867897510528564, "observed_direction": "not lower", "confirmed": false } }