{ "bench_version": 1, "track": "router_regime_regression", "model": "custom_routed_experts", "metric_direction": "lower is better", "n_seeds": 8, "baseline": { "best_cfg": { "lr": 0.01, "balance_weight": 0.0 }, "sweep": [ { "cfg": { "lr": 0.001, "balance_weight": 0.0 }, "mean": 1.2769302129745483 }, { "cfg": { "lr": 0.001, "balance_weight": 0.01 }, "mean": 1.2769302129745483 }, { "cfg": { "lr": 0.001, "balance_weight": 0.1 }, "mean": 1.2769302129745483 }, { "cfg": { "lr": 0.003, "balance_weight": 0.0 }, "mean": 1.0836826115846634 }, { "cfg": { "lr": 0.003, "balance_weight": 0.01 }, "mean": 1.0836826115846634 }, { "cfg": { "lr": 0.003, "balance_weight": 0.1 }, "mean": 1.0836826115846634 }, { "cfg": { "lr": 0.01, "balance_weight": 0.0 }, "mean": 0.5929363369941711 }, { "cfg": { "lr": 0.01, "balance_weight": 0.01 }, "mean": 0.5929363369941711 }, { "cfg": { "lr": 0.01, "balance_weight": 0.1 }, "mean": 0.5929363369941711 } ], "full": { "per_seed": [ 0.6337619423866272, 0.5049523711204529, 0.5780711770057678, 0.6549598574638367, 0.7300989627838135, 0.5037400126457214, 0.5650720000267029, 0.5289599895477295 ], "mean": 0.5874520391225815, "std": 0.08006617146577943, "details": [ { "metric": 0.6337619423866272, "history": [ 1.4282171487808228, 1.190228533744812, 1.092739050388336, 1.0344352579116822, 0.9784821963310242, 0.9456612777709961, 0.8985520505905151, 0.8510248947143555, 0.8331455397605896, 0.7536739921569824, 0.6857570123672485, 0.6390636825561523, 0.5909974932670593, 0.5384618306159973, 0.4952736568450928, 0.45394895553588865, 0.4213700973987579, 0.38855039596557617 ], "counts": [ 12.0, 119.0, 215.0, 54.0 ], "prices": [ 0.0, 0.0, 0.0, 0.0 ], "dual_gap": 36.18646240234375, "price_excess_corr": 0.0 }, { "metric": 0.5049523711204529, "history": [ 1.5119562721252442, 1.1794532775878905, 1.0590163993835449, 0.9910790395736694, 0.9429091310501099, 0.8942043042182922, 0.8424692082405091, 0.815048189163208, 0.7707590198516846, 0.7232303953170777, 0.6701731777191162, 0.6477069449424744, 0.6097545051574706, 0.5813015913963318, 0.551750590801239, 0.5354827451705932, 0.49303611755371096, 0.4774340856075287 ], "counts": [ 122.0, 67.0, 88.0, 123.0 ], "prices": [ 0.0, 0.0, 0.0, 0.0 ], "dual_gap": -3.1641159057617188, "price_excess_corr": 0.0 }, { "metric": 0.5780711770057678, "history": [ 1.818830304145813, 1.3692575263977051, 1.3353008460998534, 1.2090500450134278, 1.153546187877655, 1.094208278656006, 1.0585458588600158, 1.0022277736663818, 0.9496053886413575, 0.9010961532592774, 0.8399820518493653, 0.7756313908100129, 0.7183408498764038, 0.6691022443771363, 0.6206990718841553, 0.5716083288192749, 0.5291094565391541, 0.5017254376411437 ], "counts": [ 38.0, 97.0, 37.0, 228.0 ], "prices": [ 0.0, 0.0, 0.0, 0.0 ], "dual_gap": 20.31761932373047, "price_excess_corr": 0.0 }, { "metric": 0.6549598574638367, "history": [ 1.8027701520919799, 1.3253757810592652, 1.2136915063858031, 1.1932248497009277, 1.1371701192855834, 1.071843571662903, 1.003528470993042, 0.9638576173782348, 0.8964797353744507, 0.8383043766021728, 0.7723504877090455, 0.7259118270874023, 0.6827260518074035, 0.6220569729804992, 0.5784211730957032, 0.5290492010116578, 0.4989316189289093, 0.4740072989463806 ], "counts": [ 31.0, 161.0, 54.0, 154.0 ], "prices": [ 0.0, 0.0, 0.0, 0.0 ], "dual_gap": -5.1824951171875, "price_excess_corr": 0.0 }, { "metric": 0.7300989627838135, "history": [ 1.599121561050415, 1.2852731895446778, 1.2019276189804078, 1.1541751861572265, 1.1143722343444824, 1.0603200483322144, 1.0094259405136108, 0.9526594161987305, 0.8896933722496033, 0.8333265924453735, 0.7677580070495605, 0.7211516761779785, 0.6531311893463134, 0.6131453013420105, 0.5731891989707947, 0.5463045930862427, 0.503038067817688, 0.5016564512252808 ], "counts": [ 178.0, 168.0, 5.0, 49.0 ], "prices": [ 0.0, 0.0, 0.0, 0.0 ], "dual_gap": 0.041805267333984375, "price_excess_corr": 0.0 }, { "metric": 0.5037400126457214, "history": [ 1.6408466291427612, 1.4240649843215942, 1.3136035633087157, 1.229137110710144, 1.1278575944900513, 1.026499638557434, 0.9228047466278076, 0.8109598052501679, 0.7066186809539795, 0.6022843074798584, 0.5258250689506531, 0.49430953979492187, 0.4652757954597473, 0.44274723052978515, 0.4237631869316101, 0.39690431773662566, 0.39293508052825926, 0.38945595383644105 ], "counts": [ 0.0, 47.0, 353.0, 0.0 ], "prices": [ 0.0, 0.0, 0.0, 0.0 ], "dual_gap": 59.741790771484375, "price_excess_corr": 0.0 }, { "metric": 0.5650720000267029, "history": [ 1.2339717054367065, 0.9556390428543091, 0.8510150122642517, 0.7696062850952149, 0.7309779691696167, 0.6998401021957398, 0.6714158976078033, 0.6412617421150207, 0.6248714089393616, 0.6193566513061524, 0.5911355257034302, 0.5783848667144775, 0.5490110605955124, 0.5357205724716186, 0.5112077903747558, 0.5009259700775146, 0.470411297082901, 0.4618472504615784 ], "counts": [ 142.0, 141.0, 105.0, 12.0 ], "prices": [ 0.0, 0.0, 0.0, 0.0 ], "dual_gap": 14.618988037109375, "price_excess_corr": 0.0 }, { "metric": 0.5289599895477295, "history": [ 1.445556445121765, 1.2772163200378417, 1.2140379071235656, 1.1260969638824463, 1.0943745374679565, 1.0246259737014771, 0.9652957391738891, 0.8634569787979126, 0.7870700311660767, 0.7156670355796814, 0.6199389231204987, 0.5716872358322144, 0.5030802476406098, 0.4635558235645294, 0.41914095878601076, 0.38848516941070554, 0.36908327221870424, 0.34468125343322753 ], "counts": [ 22.0, 1.0, 355.0, 22.0 ], "prices": [ 0.0, 0.0, 0.0, 0.0 ], "dual_gap": 85.62846374511719, "price_excess_corr": 0.0 } ] } }, "idea": { "per_seed": [ 0.6441643238067627, 0.5522328615188599, 0.6585541367530823, 0.6564445495605469, 0.738909900188446, 0.6927487850189209, 0.6766970753669739, 0.6938974857330322 ], "mean": 0.6642061397433281, "std": 0.05406981038533268, "details": [ { "metric": 0.6441643238067627, "history": [ 1.4031563186645508, 1.1396907997131347, 1.030085563659668, 0.9904113817214966, 0.9025879764556884, 0.8670305633544921, 0.8303156185150147, 0.7986777687072754, 0.7985850524902344, 0.7339662647247315, 0.698540813922882, 0.644290542602539, 0.6320906519889832, 0.5695153284072876, 0.5462218189239502, 0.5224982738494873, 0.4766631841659546, 0.4712458038330078 ], "counts": [ 81.0, 114.0, 107.0, 98.0 ], "prices": [ 0.001999999862164259, 0.1849999874830246, 0.3819999098777771, 0.11400001496076584 ], "dual_gap": -0.86676025390625, "price_excess_corr": 0.7315175533294678 }, { "metric": 0.5522328615188599, "history": [ 1.5037132596969605, 1.1599038314819337, 1.046072883605957, 0.9561529111862183, 0.9125625324249268, 0.8491886901855469, 0.7865666329860688, 0.7643450164794922, 0.7428707802295684, 0.6975994086265564, 0.6477171063423157, 0.6253858184814454, 0.6086556792259217, 0.5550899171829223, 0.527807469367981, 0.5323403787612915, 0.5048370456695557, 0.4717504811286926 ], "counts": [ 92.0, 96.0, 100.0, 112.0 ], "prices": [ 0.07299996167421341, 0.005000000819563866, 0.01600000075995922, 0.05800000950694084 ], "dual_gap": -0.9556369781494141, "price_excess_corr": 0.10869929194450378 }, { "metric": 0.6585541367530823, "history": [ 1.761208825111389, 1.1728641319274902, 1.1203809022903441, 0.9686005115509033, 0.9359130239486695, 0.8751321792602539, 0.8234063625335694, 0.8007956433296204, 0.7692797470092774, 0.7613382744789123, 0.7146817946434021, 0.699941680431366, 0.6649033844470977, 0.6361303853988648, 0.6227840328216553, 0.6006997179985046, 0.56922039270401, 0.5525556755065918 ], "counts": [ 95.0, 110.0, 79.0, 116.0 ], "prices": [ 0.008000001311302185, 0.08899999409914017, 0.003999999258667231, 0.2290000021457672 ], "dual_gap": -2.62347412109375, "price_excess_corr": 0.8477886915206909 }, { "metric": 0.6564445495605469, "history": [ 1.7998930931091308, 1.2915659427642823, 1.1516711568832398, 1.1157651948928833, 1.0996507930755615, 1.0418196868896485, 0.9867689752578735, 0.9761204719543457, 0.9016803455352783, 0.87863596200943, 0.8343607997894287, 0.828931360244751, 0.765202465057373, 0.7308063387870789, 0.6686744666099549, 0.634844012260437, 0.5963686537742615, 0.5574372768402099 ], "counts": [ 86.0, 108.0, 99.0, 107.0 ], "prices": [ 0.053999993950128555, 0.13999998569488525, 2.3283064365386963e-10, 0.1860000044107437 ], "dual_gap": -4.817523956298828, "price_excess_corr": 0.6527703404426575 }, { "metric": 0.738909900188446, "history": [ 1.5758513450622558, 1.1992204332351684, 1.0975655937194824, 1.106093807220459, 1.0530373549461365, 0.9710775756835938, 0.9341824054718018, 0.8848945617675781, 0.8314888620376587, 0.8328626918792724, 0.7714112186431885, 0.7077824592590332, 0.6948019361495972, 0.6471130168437957, 0.6198636960983276, 0.5824776190519333, 0.5528994226455688, 0.5487739419937134 ], "counts": [ 81.0, 92.0, 90.0, 137.0 ], "prices": [ 0.3009999692440033, 0.31300005316734314, 0.0, 0.09099999070167542 ], "dual_gap": 0.3594779968261719, "price_excess_corr": -0.414129376411438 }, { "metric": 0.6927487850189209, "history": [ 1.578282914161682, 1.234548954963684, 1.0721720218658448, 1.0114439725875854, 0.9600631952285766, 0.9064272737503052, 0.8845085048675537, 0.8476518750190735, 0.7936444997787475, 0.7221487879753112, 0.675731315612793, 0.6356342124938965, 0.5966529107093811, 0.5677075898647308, 0.542996826171875, 0.5033763444423676, 0.47669029474258423, 0.46219895362854 ], "counts": [ 80.0, 120.0, 111.0, 89.0 ], "prices": [ 0.0, 0.40799999237060547, 0.6180000305175781, 0.08100000023841858 ], "dual_gap": -11.450798034667969, "price_excess_corr": 0.874034583568573 }, { "metric": 0.6766970753669739, "history": [ 1.2320793676376343, 0.9413762331008911, 0.8169021797180176, 0.7195290040969848, 0.691906635761261, 0.6519558382034302, 0.6347930026054383, 0.6266996812820435, 0.6043409276008606, 0.6051623976230621, 0.5799094390869141, 0.5754379844665527, 0.5715164792537689, 0.5350594699382782, 0.522064962387085, 0.5143913173675537, 0.4961080849170685, 0.512918781042099 ], "counts": [ 131.0, 106.0, 83.0, 80.0 ], "prices": [ 0.2599999010562897, 0.3229999244213104, 0.3390001356601715, 0.0 ], "dual_gap": 2.3051986694335938, "price_excess_corr": 0.3783074617385864 }, { "metric": 0.6938974857330322, "history": [ 1.3202440786361693, 1.056475853919983, 0.9645787239074707, 0.809308934211731, 0.7542427253723144, 0.694851930141449, 0.6566034412384033, 0.5901383829116821, 0.6083235359191894, 0.6138130974769592, 0.6434554958343506, 0.6685442996025085, 0.6232259523868561, 0.5891516327857971, 0.5718476819992065, 0.5396019458770752, 0.5289167666435242, 0.4906179118156433 ], "counts": [ 81.0, 95.0, 129.0, 95.0 ], "prices": [ 0.09799998253583908, 0.0, 0.5430000424385071, 0.1929999589920044 ], "dual_gap": -0.504119873046875, "price_excess_corr": 0.891348659992218 } ], "best_cfg": { "lr": 0.01, "rho": 0.001 } }, "comparison": { "delta_mean": 0.07675410062074661, "idea_wins": 0, "n_pairs": 8, "per_seed_diffs": [ 0.010402381420135498, 0.04728049039840698, 0.08048295974731445, 0.001484692096710205, 0.008810937404632568, 0.18900877237319946, 0.111625075340271, 0.16493749618530273 ], "p_value": 0.0081, "mde": 0.06092237367016021, "mde_rel_pct": 10.37061234158858, "verdict": "idea worse (significant)", "system_worked": false }, "mechanism_signature": { "mechanism_signature": { "prediction": "positive expert prices should correspond to positive realized load excess and reduce load imbalance", "trained_price_excess_correlations": [ 0.7315175533294678, 0.10869929194450378, 0.8477886915206909, 0.6527703404426575, -0.414129376411438, 0.874034583568573, 0.3783074617385864, 0.891348659992218 ], "trained_dual_gaps": [ -0.86676025390625, -0.9556369781494141, -2.62347412109375, -4.817523956298828, 0.3594779968261719, -11.450798034667969, 2.3051986694335938, -0.504119873046875 ], "mean_abs_dual_gap": 2.9853737354278564, "confirmed": true }, "idea_sweep": [ { "cfg": { "lr": 0.01, "rho": 0.001 }, "result": { "per_seed": [ 0.6441643238067627, 0.5522328615188599, 0.6585541367530823, 0.6564445495605469, 0.738909900188446, 0.6927487850189209, 0.6766970753669739, 0.6938974857330322 ], "mean": 0.6642061397433281, "std": 0.05406981038533268, "details": [ { "metric": 0.6441643238067627, "history": [ 1.4031563186645508, 1.1396907997131347, 1.030085563659668, 0.9904113817214966, 0.9025879764556884, 0.8670305633544921, 0.8303156185150147, 0.7986777687072754, 0.7985850524902344, 0.7339662647247315, 0.698540813922882, 0.644290542602539, 0.6320906519889832, 0.5695153284072876, 0.5462218189239502, 0.5224982738494873, 0.4766631841659546, 0.4712458038330078 ], "counts": [ 81.0, 114.0, 107.0, 98.0 ], "prices": [ 0.001999999862164259, 0.1849999874830246, 0.3819999098777771, 0.11400001496076584 ], "dual_gap": -0.86676025390625, "price_excess_corr": 0.7315175533294678 }, { "metric": 0.5522328615188599, "history": [ 1.5037132596969605, 1.1599038314819337, 1.046072883605957, 0.9561529111862183, 0.9125625324249268, 0.8491886901855469, 0.7865666329860688, 0.7643450164794922, 0.7428707802295684, 0.6975994086265564, 0.6477171063423157, 0.6253858184814454, 0.6086556792259217, 0.5550899171829223, 0.527807469367981, 0.5323403787612915, 0.5048370456695557, 0.4717504811286926 ], "counts": [ 92.0, 96.0, 100.0, 112.0 ], "prices": [ 0.07299996167421341, 0.005000000819563866, 0.01600000075995922, 0.05800000950694084 ], "dual_gap": -0.9556369781494141, "price_excess_corr": 0.10869929194450378 }, { "metric": 0.6585541367530823, "history": [ 1.761208825111389, 1.1728641319274902, 1.1203809022903441, 0.9686005115509033, 0.9359130239486695, 0.8751321792602539, 0.8234063625335694, 0.8007956433296204, 0.7692797470092774, 0.7613382744789123, 0.7146817946434021, 0.699941680431366, 0.6649033844470977, 0.6361303853988648, 0.6227840328216553, 0.6006997179985046, 0.56922039270401, 0.5525556755065918 ], "counts": [ 95.0, 110.0, 79.0, 116.0 ], "prices": [ 0.008000001311302185, 0.08899999409914017, 0.003999999258667231, 0.2290000021457672 ], "dual_gap": -2.62347412109375, "price_excess_corr": 0.8477886915206909 }, { "metric": 0.6564445495605469, "history": [ 1.7998930931091308, 1.2915659427642823, 1.1516711568832398, 1.1157651948928833, 1.0996507930755615, 1.0418196868896485, 0.9867689752578735, 0.9761204719543457, 0.9016803455352783, 0.87863596200943, 0.8343607997894287, 0.828931360244751, 0.765202465057373, 0.7308063387870789, 0.6686744666099549, 0.634844012260437, 0.5963686537742615, 0.5574372768402099 ], "counts": [ 86.0, 108.0, 99.0, 107.0 ], "prices": [ 0.053999993950128555, 0.13999998569488525, 2.3283064365386963e-10, 0.1860000044107437 ], "dual_gap": -4.817523956298828, "price_excess_corr": 0.6527703404426575 }, { "metric": 0.738909900188446, "history": [ 1.5758513450622558, 1.1992204332351684, 1.0975655937194824, 1.106093807220459, 1.0530373549461365, 0.9710775756835938, 0.9341824054718018, 0.8848945617675781, 0.8314888620376587, 0.8328626918792724, 0.7714112186431885, 0.7077824592590332, 0.6948019361495972, 0.6471130168437957, 0.6198636960983276, 0.5824776190519333, 0.5528994226455688, 0.5487739419937134 ], "counts": [ 81.0, 92.0, 90.0, 137.0 ], "prices": [ 0.3009999692440033, 0.31300005316734314, 0.0, 0.09099999070167542 ], "dual_gap": 0.3594779968261719, "price_excess_corr": -0.414129376411438 }, { "metric": 0.6927487850189209, "history": [ 1.578282914161682, 1.234548954963684, 1.0721720218658448, 1.0114439725875854, 0.9600631952285766, 0.9064272737503052, 0.8845085048675537, 0.8476518750190735, 0.7936444997787475, 0.7221487879753112, 0.675731315612793, 0.6356342124938965, 0.5966529107093811, 0.5677075898647308, 0.542996826171875, 0.5033763444423676, 0.47669029474258423, 0.46219895362854 ], "counts": [ 80.0, 120.0, 111.0, 89.0 ], "prices": [ 0.0, 0.40799999237060547, 0.6180000305175781, 0.08100000023841858 ], "dual_gap": -11.450798034667969, "price_excess_corr": 0.874034583568573 }, { "metric": 0.6766970753669739, "history": [ 1.2320793676376343, 0.9413762331008911, 0.8169021797180176, 0.7195290040969848, 0.691906635761261, 0.6519558382034302, 0.6347930026054383, 0.6266996812820435, 0.6043409276008606, 0.6051623976230621, 0.5799094390869141, 0.5754379844665527, 0.5715164792537689, 0.5350594699382782, 0.522064962387085, 0.5143913173675537, 0.4961080849170685, 0.512918781042099 ], "counts": [ 131.0, 106.0, 83.0, 80.0 ], "prices": [ 0.2599999010562897, 0.3229999244213104, 0.3390001356601715, 0.0 ], "dual_gap": 2.3051986694335938, "price_excess_corr": 0.3783074617385864 }, { "metric": 0.6938974857330322, "history": [ 1.3202440786361693, 1.056475853919983, 0.9645787239074707, 0.809308934211731, 0.7542427253723144, 0.694851930141449, 0.6566034412384033, 0.5901383829116821, 0.6083235359191894, 0.6138130974769592, 0.6434554958343506, 0.6685442996025085, 0.6232259523868561, 0.5891516327857971, 0.5718476819992065, 0.5396019458770752, 0.5289167666435242, 0.4906179118156433 ], "counts": [ 81.0, 95.0, 129.0, 95.0 ], "prices": [ 0.09799998253583908, 0.0, 0.5430000424385071, 0.1929999589920044 ], "dual_gap": -0.504119873046875, "price_excess_corr": 0.891348659992218 } ], "best_cfg": { "lr": 0.01, "rho": 0.001 } } }, { "cfg": { "lr": 0.01, "rho": 0.01 }, "result": { "per_seed": [ 0.6262717843055725, 0.6611071825027466, 0.7047536969184875, 0.6719289422035217, 0.7768282890319824, 0.6908259391784668, 0.6206275224685669, 0.7099313735961914 ], "mean": 0.682784341275692, "std": 0.050389091050960635, "details": [ { "metric": 0.6262717843055725, "history": [ 1.3039574146270752, 1.1375168132781983, 0.9833345127105713, 1.0023479652404785, 0.9165901398658752, 0.853428463935852, 0.828366084098816, 0.7824184989929199, 0.730939371585846, 0.7010899758338929, 0.6970420432090759, 0.619460997581482, 0.6214600610733032, 0.6054762434959412, 0.5303078866004944, 0.4873678207397461, 0.4333805441856384, 0.47833149433135985 ], "counts": [ 87.0, 120.0, 103.0, 90.0 ], "prices": [ 0.04999999701976776, 0.2099999338388443, 0.45999982953071594, 0.18000000715255737 ], "dual_gap": 0.7860870361328125, "price_excess_corr": 0.4049873650074005 }, { "metric": 0.6611071825027466, "history": [ 1.525277624130249, 1.2126392173767089, 1.0078798913955689, 1.0071892642974853, 0.9479112243652343, 0.9446576690673828, 0.8098049712181091, 0.9090130233764648, 0.8573304033279419, 0.7679864072799683, 0.7141460466384888, 0.6927071332931518, 0.6666260385513305, 0.6555062329769135, 0.6064930319786072, 0.6451085591316223, 0.6082721376419067, 0.6090406477451324 ], "counts": [ 97.0, 83.0, 145.0, 75.0 ], "prices": [ 0.1499999314546585, 0.09000000357627869, 0.030000029131770134, 0.16999995708465576 ], "dual_gap": -1.5525569915771484, "price_excess_corr": -0.8204752206802368 }, { "metric": 0.7047536969184875, "history": [ 1.7032924365997315, 1.1933565044403076, 1.1769169855117798, 0.9702837872505188, 0.9021575140953064, 0.9135065031051636, 0.8400838613510132, 0.8087026262283326, 0.76284184217453, 0.7864427018165588, 0.7398368334770202, 0.6987604725360871, 0.7155058979988098, 0.7891936492919922, 0.6775259113311768, 0.6355217003822327, 0.6297998356819153, 0.6280114138126374 ], "counts": [ 104.0, 113.0, 77.0, 106.0 ], "prices": [ 0.0599999725818634, 0.13999998569488525, 0.07999999076128006, 0.3200000822544098 ], "dual_gap": -3.084991455078125, "price_excess_corr": 0.3812926113605499 }, { "metric": 0.6719289422035217, "history": [ 1.7835893440246582, 1.3398815345764161, 1.1745865392684935, 1.1184628486633301, 1.0979045367240905, 1.087763638496399, 1.085881280899048, 1.0113541078567505, 0.9435374069213868, 0.938734905719757, 0.922377724647522, 0.8435751342773438, 0.7949281311035157, 0.7932442426681519, 0.7761956405639648, 0.7285706949234009, 0.698066930770874, 0.67836585521698 ], "counts": [ 126.0, 56.0, 113.0, 105.0 ], "prices": [ 0.09999997168779373, 0.24999991059303284, 0.019999999552965164, 0.24999995529651642 ], "dual_gap": 0.5536575317382812, "price_excess_corr": -0.6561136245727539 }, { "metric": 0.7768282890319824, "history": [ 1.4601379203796387, 1.1082266092300415, 1.1362065076828003, 1.0652369022369386, 0.9823706912994384, 0.9776483631134033, 0.9667921781539917, 0.9053830432891846, 0.9509675192832947, 0.8583141326904297, 0.8127024602890015, 0.8199803376197815, 0.7261572551727294, 0.7076331377029419, 0.762781138420105, 0.6751837837696075, 0.6537091398239135, 0.6659013748168945 ], "counts": [ 117.0, 95.0, 86.0, 102.0 ], "prices": [ 0.3199998140335083, 0.3899996280670166, 0.0599999874830246, 0.2099999189376831 ], "dual_gap": -3.830963134765625, "price_excess_corr": 0.5433895587921143 }, { "metric": 0.6908259391784668, "history": [ 1.6264865016937256, 1.1661974811553955, 0.9919495415687561, 0.9857214403152466, 0.8792146921157837, 0.9436957502365112, 0.9131818580627441, 0.7837913727760315, 0.8391086673736572, 0.7410305190086365, 0.67766282081604, 0.6405617022514343, 0.654193127155304, 0.5937393045425415, 0.5505223798751832, 0.48757129430770874, 0.47665258169174196, 0.5111252760887146 ], "counts": [ 105.0, 91.0, 122.0, 82.0 ], "prices": [ 0.029999995604157448, 0.5300000905990601, 0.6700002551078796, 0.1600000113248825 ], "dual_gap": -4.6856689453125, "price_excess_corr": 0.4581056237220764 }, { "metric": 0.6206275224685669, "history": [ 1.2146726608276368, 0.8417455530166627, 0.8210588359832763, 0.8491046547889709, 0.7547090137004853, 0.7639045429229736, 0.705108392238617, 0.693201768398285, 0.6587955093383789, 0.6621153998374939, 0.6825254893302918, 0.6306395292282104, 0.6202736258506775, 0.6084072184562683, 0.5361759781837463, 0.6167222261428833, 0.5254430556297303, 0.49973246693611145 ], "counts": [ 98.0, 82.0, 92.0, 128.0 ], "prices": [ 0.37999996542930603, 0.4399999976158142, 0.41999995708465576, 0.030000001192092896 ], "dual_gap": 6.8090972900390625, "price_excess_corr": -0.9755735397338867 }, { "metric": 0.7099313735961914, "history": [ 1.273847484588623, 0.9680283117294312, 0.9676177501678467, 0.7686456203460693, 0.744644775390625, 0.8059906673431396, 0.7485140085220336, 0.692188173532486, 0.8139085674285889, 0.8050192737579346, 0.7376338243484497, 0.6645472645759583, 0.6088972043991089, 0.5972435641288757, 0.5730879020690918, 0.6306612575054169, 0.5813390743732453, 0.47985951900482177 ], "counts": [ 95.0, 113.0, 76.0, 116.0 ], "prices": [ 0.13999997079372406, 0.039999984204769135, 0.6700000762939453, 0.23999977111816406 ], "dual_gap": 11.904136657714844, "price_excess_corr": -0.8073238730430603 } ] } }, { "cfg": { "lr": 0.01, "rho": 0.1 }, "result": { "per_seed": [ 1.1683597564697266, 0.9057276248931885, 0.9819811582565308, 1.0769124031066895, 1.124353051185608, 0.979142427444458, 1.2250404357910156, 1.028434157371521 ], "mean": 1.0612438768148422, "std": 0.10744859492285688, "details": [ { "metric": 1.1683597564697266, "history": [ 1.4490776538848877, 1.5859220504760743, 1.3607615947723388, 1.4989393043518067, 1.3760244274139404, 1.3536517190933228, 1.3585372638702393, 1.3083824491500855, 1.2400494956970214, 1.3978913164138793, 1.268368067741394, 1.2482490491867066, 1.2231985998153687, 1.1448024535179138, 1.1257777547836303, 1.0693637561798095, 0.9765922141075134, 1.0164781403541565 ], "counts": [ 0.0, 0.0, 321.0, 79.0 ], "prices": [ 5.100002765655518, 4.300004005432129, 1.3000054359436035, 1.3000035285949707 ], "dual_gap": 711.1893463134766, "price_excess_corr": -0.7495322823524475 }, { "metric": 0.9057276248931885, "history": [ 1.6662955570220948, 1.6366234970092775, 1.5101047563552856, 1.5987392616271974, 1.4745655584335327, 1.4979643392562867, 1.3766446256637572, 1.3513256263732911, 1.3477717399597169, 1.3279736328125, 1.2432228183746339, 1.2315861248970033, 1.1717516899108886, 1.0781730556488036, 1.0162011313438415, 0.9924906349182129, 0.8533278799057007, 0.8498016858100891 ], "counts": [ 92.0, 0.0, 0.0, 308.0 ], "prices": [ 2.300004005432129, 3.900001049041748, 4.60000467300415, 2.100005626678467 ], "dual_gap": 400.21055603027344, "price_excess_corr": -0.8114298582077026 }, { "metric": 0.9819811582565308, "history": [ 1.8330887985229491, 1.7522169303894044, 1.7377414941787719, 1.614824047088623, 1.5683921241760255, 1.5938614463806153, 1.5247077465057373, 1.5466945004463195, 1.404871940612793, 1.451461374759674, 1.3511654186248778, 1.3450501358509064, 1.2249249458312987, 1.2138253259658813, 1.1642186546325684, 1.0807652282714844, 0.9764101076126098, 0.971161847114563 ], "counts": [ 110.0, 290.0, 0.0, 0.0 ], "prices": [ 2.0000011920928955, 1.8000009059906006, 4.000002861022949, 5.100006580352783 ], "dual_gap": 533.8030090332031, "price_excess_corr": -0.8358966112136841 }, { "metric": 1.0769124031066895, "history": [ 1.7902608394622803, 1.4834549045562744, 1.5026324462890626, 1.3873037981987, 1.426707820892334, 1.3190795135498048, 1.2909542369842528, 1.3072921562194824, 1.1851996898651123, 1.2980601835250853, 1.1551457309722901, 1.1571266746520996, 1.0714560890197753, 1.176503300666809, 1.0852110576629639, 1.0225802326202393, 0.9678384566307068, 0.9272657442092895 ], "counts": [ 0.0, 3.0, 397.0, 0.0 ], "prices": [ 2.900002956390381, 2.8000073432922363, 1.5000014305114746, 5.800005912780762 ], "dual_gap": 567.5848979949951, "price_excess_corr": -0.6454119682312012 }, { "metric": 1.124353051185608, "history": [ 1.7129732704162597, 1.7161536598205567, 1.6164196014404297, 1.5220919799804689, 1.5200805282592773, 1.4736564350128174, 1.5009299993515015, 1.4522486782073976, 1.4154387760162352, 1.405643253326416, 1.3363902616500853, 1.2636822080612182, 1.3516197395324707, 1.2489344120025634, 1.1715259742736817, 1.1471354389190673, 1.05180935382843, 1.0683834600448607 ], "counts": [ 0.0, 8.0, 392.0, 0.0 ], "prices": [ 4.000000953674316, 2.30000638961792, 1.200000286102295, 4.800002098083496 ], "dual_gap": 618.683934211731, "price_excess_corr": -0.7796272039413452 }, { "metric": 0.979142427444458, "history": [ 1.7750358438491822, 1.5844729232788086, 1.6429149341583251, 1.5185857009887695, 1.502098972797394, 1.544996886253357, 1.422248454093933, 1.4559745836257934, 1.4031295108795165, 1.3522978973388673, 1.300237669944763, 1.2396622562408448, 1.217160804271698, 1.1100136756896972, 1.1584144115447998, 0.981019310951233, 1.005128688812256, 0.8883059167861939 ], "counts": [ 0.0, 2.0, 398.0, 0.0 ], "prices": [ 4.800004005432129, 2.2000069618225098, 1.9000087976455688, 4.000002861022949 ], "dual_gap": 597.8818855285645, "price_excess_corr": -0.6340560913085938 }, { "metric": 1.2250404357910156, "history": [ 1.3113063716888427, 1.2002628803253175, 1.4514206790924071, 1.330623984336853, 1.232826681137085, 1.2343617129325866, 1.2105679941177367, 1.170863471031189, 1.1571433877944945, 1.1489333415031433, 1.1103823566436768, 1.1372274017333985, 1.0873634719848633, 1.0486375761032105, 0.9828113007545471, 1.0502234315872192, 0.9858989048004151, 0.9104355192184448 ], "counts": [ 0.0, 400.0, 0.0, 0.0 ], "prices": [ 3.600005626678467, 1.30000638961792, 5.700007438659668, 2.0000052452087402 ], "dual_gap": 776.7353515625, "price_excess_corr": -0.6313034892082214 }, { "metric": 1.028434157371521, "history": [ 1.684798412322998, 1.5274455451965332, 1.4914986324310302, 1.3847495126724243, 1.5005753707885743, 1.4229538440704346, 1.40796156167984, 1.3940145444869996, 1.3152224588394166, 1.2695887804031372, 1.2292287492752074, 1.0851865339279174, 1.1410926818847655, 1.0864189529418946, 0.9851797819137573, 0.9458699417114258, 0.834864010810852, 0.8689987945556641 ], "counts": [ 0.0, 400.0, 0.0, 0.0 ], "prices": [ 1.9000014066696167, 0.6000056266784668, 4.200000286102295, 6.000000953674316 ], "dual_gap": 983.6130700111389, "price_excess_corr": -0.7151221632957458 } ] } } ], "custom_track": { "name": "router_regime_regression", "file": "bench/custom_tracks/router_regime_regression.py", "domain": "moe-routing" } }, "paired_raw_diffs": [ 0.010402381420135498, 0.04728049039840698, 0.08048295974731445, 0.001484692096710205, 0.008810937404632568, 0.18900877237319946, 0.111625075340271, 0.16493749618530273 ] }