{ "bench_version": 1, "track": "dynamics", "model": "rnn_small", "metric_direction": "lower is better", "n_seeds": 8, "baseline": { "best_cfg": { "lr": 0.01, "pressure": 0.2 }, "sweep": [ { "cfg": { "lr": 0.001, "pressure": 0.0 }, "mean": 0.0035706528869923204 }, { "cfg": { "lr": 0.001, "pressure": 0.05 }, "mean": 0.003419904154725373 }, { "cfg": { "lr": 0.001, "pressure": 0.2 }, "mean": 0.0036505938041955233 }, { "cfg": { "lr": 0.003, "pressure": 0.0 }, "mean": 0.005002945079468191 }, { "cfg": { "lr": 0.003, "pressure": 0.05 }, "mean": 0.005080707313027233 }, { "cfg": { "lr": 0.003, "pressure": 0.2 }, "mean": 0.005320765834767371 }, { "cfg": { "lr": 0.01, "pressure": 0.0 }, "mean": 0.0013175007770769298 }, { "cfg": { "lr": 0.01, "pressure": 0.05 }, "mean": 0.0012840094132116064 }, { "cfg": { "lr": 0.01, "pressure": 0.2 }, "mean": 0.0011667622748063877 } ], "full": { "mean": 0.0011755235827877186, "std": 0.0003522616089917298, "per_seed": [ 0.0008392606396228075, 0.0014225966297090054, 0.0006899926229380071, 0.001715199206955731, 0.0013351935194805264, 0.0015198795590549707, 0.0010780569864436984, 0.0008040094980970025 ], "n": 8 } }, "idea": { "mean": 0.0010581257738522254, "std": 0.0002584312332048161, "per_seed": [ 0.0007879937184043229, 0.0010015293955802917, 0.0013035451993346214, 0.001240157289430499, 0.0013518166961148381, 0.0013151094317436218, 0.0007202015840448439, 0.0007446528761647642 ], "n": 8, "details": [ { "I_train": 0.6254459023475647, "I_test": 0.8024823665618896, "lambda": 0.5281313902530512, "H_goal": 0.55, "pressure": 0.2 }, { "I_train": 0.8220813274383545, "I_test": 1.0823215246200562, "lambda": 0.0, "H_goal": 0.55, "pressure": 0.2 }, { "I_train": 0.6989840865135193, "I_test": 0.7982175350189209, "lambda": 0.7593800543184935, "H_goal": 0.55, "pressure": 0.2 }, { "I_train": 0.5356451272964478, "I_test": 0.6818308234214783, "lambda": 0.8704148933857491, "H_goal": 0.55, "pressure": 0.2 }, { "I_train": 0.9433308243751526, "I_test": 0.857042670249939, "lambda": 0.5193827239810556, "H_goal": 0.55, "pressure": 0.2 }, { "I_train": 0.7470368146896362, "I_test": 0.62962806224823, "lambda": 0.8655181587285236, "H_goal": 0.55, "pressure": 0.2 }, { "I_train": 0.7524755001068115, "I_test": 1.1005165576934814, "lambda": 0.0, "H_goal": 0.55, "pressure": 0.2 }, { "I_train": 0.5790018439292908, "I_test": 0.9611145853996277, "lambda": 0.40624716999978266, "H_goal": 0.55, "pressure": 0.2 } ], "best_cfg": { "lr": 0.01, "pressure": 0.2, "eta_lambda": 0.08 }, "candidates": [ { "cfg": { "lr": 0.01, "pressure": 0.2, "eta_lambda": 0.04 }, "mean": 0.001126833783928305 }, { "cfg": { "lr": 0.01, "pressure": 0.2, "eta_lambda": 0.08 }, "mean": 0.0010581257738522254 }, { "cfg": { "lr": 0.01, "pressure": 0.2, "eta_lambda": 0.16 }, "mean": 0.0011072655142925214 } ] }, "comparison": { "delta_mean": -0.00011739780893549323, "idea_wins": 6, "n_pairs": 8, "per_seed_diffs": [ -5.126692121848464e-05, -0.0004210672341287136, 0.0006135525763966143, -0.00047504191752523184, 1.6623176634311676e-05, -0.00020477012731134892, -0.0003578554023988545, -5.935662193223834e-05 ], "p_value": 0.365, "mde": 0.00029061097353965813, "mde_rel_pct": 24.721832704579434, "verdict": "no significant win", "system_worked": false }, "mechanism_signature": { "bench_report": { "track_match": "controlled dynamics / latent predictive information", "custom_track": null }, "signature": { "H_goal": 0.55, "baseline_fixed_pressure": 0.2, "baseline_mean_I_test": 0.867881640791893, "idea_mean_I_test": 0.8641442656517029, "idea_mean_lambda": 0.493634298833332, "predicted": "dual pressure increases when I_test/EMA is below H_goal and should reduce deficit", "observed_deficit_baseline": -0.31788164079189296, "observed_deficit_idea": -0.31414426565170284, "confirmed": false } } }