{ "bench_version": 1, "track": "dynamics", "model": "rnn_small", "metric_direction": "lower is better", "n_seeds": 8, "baseline": { "best_cfg": { "lr": 0.006, "weight_decay": 0.0 }, "sweep": [ { "cfg": { "lr": 0.0015, "weight_decay": 0.0 }, "mean": 0.0020976006926503032 }, { "cfg": { "lr": 0.003, "weight_decay": 0.0 }, "mean": 0.0012262430100236088 }, { "cfg": { "lr": 0.006, "weight_decay": 0.0 }, "mean": 0.0010091856674989685 } ], "full": { "mean": 0.0009940863528754562, "std": 0.0002634151189555855, "per_seed": [ 0.0007214772049337626, 0.0010892449645325541, 0.0007245641318149865, 0.001501456368714571, 0.0007053734152577817, 0.0009875509422272444, 0.001235381467267871, 0.0009876423282548785 ], "n": 8 } }, "idea": { "mean": 0.0009940863528754562, "std": 0.0002634151189555855, "per_seed": [ 0.0007214772049337626, 0.0010892449645325541, 0.0007245641318149865, 0.001501456368714571, 0.0007053734152577817, 0.0009875509422272444, 0.001235381467267871, 0.0009876423282548785 ], "n": 8 }, "comparison": { "delta_mean": 0.0, "idea_wins": 0, "n_pairs": 8, "per_seed_diffs": [ 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0 ], "p_value": 1.0, "mde": 0.0, "mde_rel_pct": 0.0, "verdict": "no measurable effect", "system_worked": false }, "mechanism_signature": { "quantity": "trained recurrent-map BT coefficients a,b from centered FD along SVD kernel direction", "predicted": "BT penalty reduces fraction of checkpoints/models with |a*b| below threshold", "threshold": 0.0004, "baseline_near_degenerate_fraction": 0.75, "idea_near_degenerate_fraction": 0.75, "per_seed": [ { "seed": 0, "baseline": { "a": -0.0035965442657470703, "b": 0.011935234069824219, "ab": 4.2925596062559634e-05, "rank_smin": 0.5489700436592102 }, "idea": { "a": -0.0035965442657470703, "b": 0.011935234069824219, "ab": 4.2925596062559634e-05, "rank_smin": 0.5489700436592102 } }, { "seed": 1, "baseline": { "a": 0.01983165740966797, "b": -0.07238030433654785, "ab": 0.0014354214072227478, "rank_smin": 0.4916006326675415 }, "idea": { "a": 0.01983165740966797, "b": -0.07238030433654785, "ab": 0.0014354214072227478, "rank_smin": 0.4916006326675415 } }, { "seed": 2, "baseline": { "a": -0.021853148937225342, "b": 0.08207321166992188, "ab": 0.001793558127246797, "rank_smin": 0.48579904437065125 }, "idea": { "a": -0.021853148937225342, "b": 0.08207321166992188, "ab": 0.001793558127246797, "rank_smin": 0.48579904437065125 } }, { "seed": 3, "baseline": { "a": -0.00858500599861145, "b": 0.031914710998535156, "ab": 0.0002739879710134119, "rank_smin": 0.4880279004573822 }, "idea": { "a": -0.00858500599861145, "b": 0.031914710998535156, "ab": 0.0002739879710134119, "rank_smin": 0.4880279004573822 } }, { "seed": 4, "baseline": { "a": 0.0015027821063995361, "b": -0.0048029422760009766, "ab": 7.217775873868959e-06, "rank_smin": 0.5201539397239685 }, "idea": { "a": 0.0015027821063995361, "b": -0.0048029422760009766, "ab": 7.217775873868959e-06, "rank_smin": 0.5201539397239685 } }, { "seed": 5, "baseline": { "a": 0.00692903995513916, "b": -0.02565741539001465, "ab": 0.00017778125766199082, "rank_smin": 0.5120712518692017 }, "idea": { "a": 0.00692903995513916, "b": -0.02565741539001465, "ab": 0.00017778125766199082, "rank_smin": 0.5120712518692017 } }, { "seed": 6, "baseline": { "a": 0.0017018616199493408, "b": -0.004662871360778809, "ab": 7.935561370686628e-06, "rank_smin": 0.4283751845359802 }, "idea": { "a": 0.0017018616199493408, "b": -0.004662871360778809, "ab": 7.935561370686628e-06, "rank_smin": 0.4283751845359802 } }, { "seed": 7, "baseline": { "a": 0.002718716859817505, "b": -0.008226633071899414, "ab": 2.2365886252373457e-05, "rank_smin": 0.5046995282173157 }, "idea": { "a": 0.002718716859817505, "b": -0.008226633071899414, "ab": 2.2365886252373457e-05, "rank_smin": 0.5046995282173157 } } ], "confirmed": false }, "idea_sweep": [ { "cfg": { "lr": 0.006, "weight_decay": 0.0 }, "mean": 0.0009940863528754562, "per_seed": [ 0.0007214772049337626, 0.0010892449645325541, 0.0007245641318149865, 0.001501456368714571, 0.0007053734152577817, 0.0009875509422272444, 0.001235381467267871, 0.0009876423282548785 ] }, { "cfg": { "lr": 0.0015, "weight_decay": 0.0 }, "mean": 0.002549259254010394, "per_seed": [ 0.0012304666452109814, 0.003008471569046378, 0.001936021144501865, 0.0022154434118419886, 0.0010678971884772182, 0.0025886795483529568, 0.004030377604067326, 0.00431671692058444 ] }, { "cfg": { "lr": 0.006, "weight_decay": 0.0 }, "mean": 0.0009940863528754562, "per_seed": [ 0.0007214772049337626, 0.0010892449645325541, 0.0007245641318149865, 0.001501456368714571, 0.0007053734152577817, 0.0009875509422272444, 0.001235381467267871, 0.0009876423282548785 ] } ] }