{ "root_check": [ { "chi": 0.8, "max_real_root": 0.0330863273564016 }, { "chi": 1.0, "max_real_root": 0.07718936960152892 }, { "chi": 1.2, "max_real_root": 0.11566900160083154 } ], "bench_report": { "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": 0.42837614566087723 }, { "cfg": { "lr": 0.003 }, "mean": 0.05120387952774763 }, { "cfg": { "lr": 0.01 }, "mean": 0.0024460060812998563 } ], "full": { "mean": 0.0025610758020775393, "std": 0.0009777445230047094, "per_seed": [ 0.004299916326999664, 0.002046577399596572, 0.0017879348015412688, 0.0016495957970619202, 0.0030561184976249933, 0.003266895655542612, 0.0031689393799751997, 0.0012126285582780838 ], "n": 8 } }, "idea": { "per_seed": [ 0.004299916326999664, 0.002046577399596572, 0.0017879348015412688, 0.0016495957970619202, 0.0030561184976249933, 0.003266895655542612, 0.0031689393799751997, 0.0012126285582780838 ] }, "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": { "prediction": "online RH cap enforces chi <= rho=0.8", "predicted_max_chi": 0.8, "observed_max_chi_idea": 3.605769230769231e-06, "observed_max_chi_baseline": NaN, "observed_mean_gain_idea": 1.0000000000000002e-06, "confirmed": true } }, "idea_settings": { "0.001": [ { "metric": 0.3440668284893036, "history": [ 0.6993419528007507, 0.6813979518413543, 0.650027756690979, 0.6101660060882569, 0.5640302371978759, 0.5151371550559998, 0.465482257604599, 0.4180481779575348 ], "max_chi": 3.605769230769231e-07, "mean_effective_lr": 0.001, "mean_observed_gain": 1.0000000000000001e-07, "max_curvature": 0.0001 }, { "metric": 0.4274647533893585, "history": [ 0.7602336311340332, 0.7395647692680359, 0.7040573501586914, 0.66268874168396, 0.6158934926986694, 0.5678855276107788, 0.5177900528907776, 0.4694719660282135 ], "max_chi": 3.605769230769231e-07, "mean_effective_lr": 0.001, "mean_observed_gain": 1.0000000000000001e-07, "max_curvature": 0.0001 }, { "metric": 0.4056340754032135, "history": [ 0.6523151779174805, 0.6336322736740112, 0.5997181987762451, 0.5587008571624756, 0.5148982381820679, 0.47241021394729615, 0.4275295877456665, 0.38603972673416137 ], "max_chi": 3.605769230769231e-07, "mean_effective_lr": 0.001, "mean_observed_gain": 1.0000000000000001e-07, "max_curvature": 0.0001 }, { "metric": 0.5363389253616333, "history": [ 0.7145644426345825, 0.6977586770057678, 0.6691723155975342, 0.631492304801941, 0.5912015914916993, 0.5492246544361115, 0.5076988697052002, 0.4674421977996826 ], "max_chi": 3.605769230769231e-07, "mean_effective_lr": 0.001, "mean_observed_gain": 1.0000000000000001e-07, "max_curvature": 0.0001 }, { "metric": 0.33050650358200073, "history": [ 0.6300368976593017, 0.6130261039733886, 0.5824478936195373, 0.5431585824489593, 0.49994396686553955, 0.4560388541221619, 0.4114214777946472, 0.3689022183418274 ], "max_chi": 3.605769230769231e-07, "mean_effective_lr": 0.001, "mean_observed_gain": 1.0000000000000001e-07, "max_curvature": 0.0001 }, { "metric": 0.4261091947555542, "history": [ 0.7953370833396911, 0.7722095775604249, 0.7340326881408692, 0.684270613193512, 0.6333953094482422, 0.5830701923370362, 0.5291142201423645, 0.4788055396080017 ], "max_chi": 3.605769230769231e-07, "mean_effective_lr": 0.001, "mean_observed_gain": 1.0000000000000001e-07, "max_curvature": 0.0001 }, { "metric": 0.37957045435905457, "history": [ 0.7479315209388733, 0.7267511749267578, 0.6888936424255371, 0.6428964591026306, 0.5938624811172485, 0.544820841550827, 0.4944131255149841, 0.4452679681777954 ], "max_chi": 3.605769230769231e-07, "mean_effective_lr": 0.001, "mean_observed_gain": 1.0000000000000001e-07, "max_curvature": 0.0001 }, { "metric": 0.4494275450706482, "history": [ 0.7529299116134643, 0.7330988574028016, 0.699141960144043, 0.6572146892547608, 0.6124954748153687, 0.5655594277381897, 0.5191149485111236, 0.4722910523414612 ], "max_chi": 3.605769230769231e-07, "mean_effective_lr": 0.001, "mean_observed_gain": 1.0000000000000001e-07, "max_curvature": 0.0001 } ], "0.003": [ { "metric": 0.018445655703544617, "history": [ 0.694959774017334, 0.6420108795166015, 0.5524764275550842, 0.44393115520477294, 0.3260509085655212, 0.21360599756240845, 0.11821861624717712, 0.052244496941566465 ], "max_chi": 1.0817307692307691e-06, "mean_effective_lr": 0.003, "mean_observed_gain": 3.0000000000000004e-07, "max_curvature": 0.0001 }, { "metric": 0.051231302320957184, "history": [ 0.7553361463546753, 0.6949287056922913, 0.5964539051055908, 0.48899065852165224, 0.375873681306839, 0.26908132374286653, 0.17072125911712646, 0.09357141584157944 ], "max_chi": 1.0817307692307691e-06, "mean_effective_lr": 0.003, "mean_observed_gain": 3.0000000000000004e-07, "max_curvature": 0.0001 }, { "metric": 0.04099540039896965, "history": [ 0.6478597497940064, 0.5935505867004395, 0.5012018966674805, 0.39846776485443114, 0.29653573989868165, 0.20853621542453765, 0.12716419875621796, 0.06803727239370345 ], "max_chi": 1.0817307692307691e-06, "mean_effective_lr": 0.003, "mean_observed_gain": 3.0000000000000004e-07, "max_curvature": 0.0001 }, { "metric": 0.09414315968751907, "history": [ 0.7108429098129272, 0.6615033912658691, 0.5811297416687011, 0.4809538173675537, 0.3801296353340149, 0.28171004056930543, 0.19256000876426696, 0.11847457110881805 ], "max_chi": 1.0817307692307691e-06, "mean_effective_lr": 0.003, "mean_observed_gain": 3.0000000000000004e-07, "max_curvature": 0.0001 }, { "metric": 0.020061630755662918, "history": [ 0.626112835407257, 0.5758749866485595, 0.4887616550922394, 0.3829389560222626, 0.2756062078475952, 0.17953147500753402, 0.10008062392473221, 0.04575570151209831 ], "max_chi": 1.0817307692307691e-06, "mean_effective_lr": 0.003, "mean_observed_gain": 3.0000000000000004e-07, "max_curvature": 0.0001 }, { "metric": 0.03951025754213333, "history": [ 0.7898440504074097, 0.7223627233505249, 0.616898854970932, 0.4893602728843689, 0.3690385365486145, 0.25974723160266877, 0.1579497426748276, 0.08233777940273285 ], "max_chi": 1.0817307692307691e-06, "mean_effective_lr": 0.003, "mean_observed_gain": 3.0000000000000004e-07, "max_curvature": 0.0001 }, { "metric": 0.031925592571496964, "history": [ 0.7429254531860352, 0.6810082459449768, 0.5761089265346527, 0.4574007022380829, 0.3400264596939087, 0.2332361841201782, 0.13982866019010542, 0.06969675615429878 ], "max_chi": 1.0817307692307691e-06, "mean_effective_lr": 0.003, "mean_observed_gain": 3.0000000000000004e-07, "max_curvature": 0.0001 }, { "metric": 0.05982330068945885, "history": [ 0.7481838035583496, 0.6902918517589569, 0.5962544202804565, 0.4881104326248169, 0.3803297543525696, 0.27569568634033204, 0.18162715554237366, 0.10307565599679946 ], "max_chi": 1.0817307692307691e-06, "mean_effective_lr": 0.003, "mean_observed_gain": 3.0000000000000004e-07, "max_curvature": 0.0001 } ], "0.01": [ { "metric": 0.004299916326999664, "history": [ 0.6798131060600281, 0.512830947637558, 0.26385428369045255, 0.055778834596276285, 0.017274612039327623, 0.07015145182609558, 0.0401259084790945, 0.0036679642274975778 ], "max_chi": 3.605769230769231e-06, "mean_effective_lr": 0.01, "mean_observed_gain": 1.0000000000000002e-06, "max_curvature": 0.0001 }, { "metric": 0.002046577399596572, "history": [ 0.7385136270523072, 0.5534801185131073, 0.2967984753847122, 0.08821958027780057, 0.009234225079417228, 0.047636431157588956, 0.04728898391127587, 0.010239683538675309 ], "max_chi": 3.605769230769231e-06, "mean_effective_lr": 0.01, "mean_observed_gain": 1.0000000000000002e-06, "max_curvature": 0.0001 }, { "metric": 0.0017879348015412688, "history": [ 0.6325315642356872, 0.469499454498291, 0.2445989990234375, 0.06959607563912869, 0.013648657500743866, 0.05180964022874832, 0.043835976719856264, 0.008696820829063654 ], "max_chi": 3.605769230769231e-06, "mean_effective_lr": 0.01, "mean_observed_gain": 1.0000000000000002e-06, "max_curvature": 0.0001 }, { "metric": 0.0016495957970619202, "history": [ 0.6980253458023071, 0.544585394859314, 0.32391603231430055, 0.10666814312338829, 0.00840164065361023, 0.041370466351509094, 0.06321462988853455, 0.022033566143363714 ], "max_chi": 3.605769230769231e-06, "mean_effective_lr": 0.01, "mean_observed_gain": 1.0000000000000002e-06, "max_curvature": 0.0001 }, { "metric": 0.0030561184976249933, "history": [ 0.6125363111495972, 0.4540520048141479, 0.21791321754455567, 0.036135574914515016, 0.023327344805002214, 0.06946323305368424, 0.03671337585896253, 0.0036678536795079707 ], "max_chi": 3.605769230769231e-06, "mean_effective_lr": 0.01, "mean_observed_gain": 1.0000000000000002e-06, "max_curvature": 0.0001 }, { "metric": 0.003266895655542612, "history": [ 0.7710019874572754, 0.5653183555603027, 0.29373787581920624, 0.06810168888419867, 0.017942358553409577, 0.058926407247781754, 0.041252076476812366, 0.006175353517755866 ], "max_chi": 3.605769230769231e-06, "mean_effective_lr": 0.01, "mean_observed_gain": 1.0000000000000002e-06, "max_curvature": 0.0001 }, { "metric": 0.0031689393799751997, "history": [ 0.7256936359405518, 0.5360211396217346, 0.2658588397502899, 0.05989395052194595, 0.02104116842150688, 0.06888968884944915, 0.047494716197252276, 0.007844956731423736 ], "max_chi": 3.605769230769231e-06, "mean_effective_lr": 0.01, "mean_observed_gain": 1.0000000000000002e-06, "max_curvature": 0.0001 }, { "metric": 0.0012126285582780838, "history": [ 0.7319039154052734, 0.554954423904419, 0.3081369322538376, 0.09451350867748261, 0.007374951243400573, 0.04413643956184387, 0.05444020733237267, 0.015588753465563059 ], "max_chi": 3.605769230769231e-06, "mean_effective_lr": 0.01, "mean_observed_gain": 1.0000000000000002e-06, "max_curvature": 0.0001 } ] }, "custom_track": null }