{ "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": 0.07048395439051092 }, { "cfg": { "lr": 0.003 }, "mean": 0.0036796721105929464 }, { "cfg": { "lr": 0.01 }, "mean": 0.0020796392345800996 } ], "full": { "mean": 0.0018650785204954445, "per_seed": [ 0.0020393452141433954, 0.0022905797231942415, 0.0019529510755091906, 0.002035680925473571, 0.0009734529303386807, 0.003012745874002576, 0.0011985950404778123, 0.001417277380824089 ] } }, "idea": { "mean": 0.07028227765113115, "per_seed": [ 0.09344271570444107, 0.035067640244960785, 0.06569516658782959, 0.06504810601472855, 0.057421158999204636, 0.045870814472436905, 0.06615367531776428, 0.1335589438676834 ] }, "comparison": { "delta_mean": 0.06841719913063571, "idea_wins": 0, "n_pairs": 8, "per_seed_diffs": [ 0.09140337049029768, 0.03277706052176654, 0.0637422155123204, 0.06301242508925498, 0.056447706068865955, 0.04285806859843433, 0.06495508027728647, 0.13214166648685932 ], "p_value": 0.0081, "mde": 0.02590531560040504, "mde_rel_pct": 1388.9664867044569, "verdict": "idea worse (significant)", "system_worked": false }, "idea_sweep": [ { "cfg": { "lr": 0.001 }, "mean": 0.6029151976108551 }, { "cfg": { "lr": 0.003 }, "mean": 0.2946254089474678 }, { "cfg": { "lr": 0.01 }, "mean": 0.06481340713799 } ], "architecture": "matched 64-unit recurrent state and scalar head; only deterministic leaky update vs correlated IF differs", "mechanism_signature": { "predicted_shared_covariance_factor": 0.42250000000000004, "observed_shared_noise_correlation_mean": 0.42043366931669834, "observed_shared_noise_correlation_per_seed": [ 0.4126678738329091, 0.4521553626790912, 0.41302410068546774, 0.40544973371858345, 0.4157424909414776, 0.44676270990572226, 0.4155898912561236, 0.40207719151421156 ], "spike_rate_mean": 0.03861450147815049, "confirmed": true } }