assay-compiled-base / eval-dev.json
Berk's picture
Upload folder using huggingface_hub
003acf8 verified
Raw
History Blame Contribute Delete
22.9 kB
{
"overall": {
"count": 6113,
"accuracy": 0.6679208244724358,
"brier": 0.4544455845096957,
"brier_soft": 0.4310818103173888,
"nll": 0.8546754183783308,
"nll_soft": 0.8720034822503786,
"ece": 0.10337806078709712,
"mean_confidence": 0.769263460145617,
"confident_error_rate": 0.05300179944380828,
"score_mae": 0.53153959018479,
"evidence_auc": 0.9074727460810058
},
"by_type": {
"bool": {
"count": 1898,
"accuracy": 0.7787144362486829,
"brier": 0.32803778696812624,
"brier_soft": 0.32512087658844907,
"nll": 0.5308636866774319,
"nll_soft": 0.5373907544543911,
"ece": 0.0827987949639653,
"mean_confidence": 0.8615132312126481,
"confident_error_rate": 0.07586933614330875,
"evidence_auc": 0.9039322111530974
},
"choice": {
"count": 3248,
"accuracy": 0.625615763546798,
"brier": 0.5167065686871394,
"brier_soft": 0.4935414744397849,
"nll": 1.034598991800825,
"nll_soft": 1.0506837306443417,
"ece": 0.13399685761808722,
"mean_confidence": 0.7569434392263682,
"confident_error_rate": 0.05511083743842365,
"evidence_auc": 0.9208606009429315
},
"score": {
"count": 967,
"accuracy": 0.5925542916235781,
"brier": 0.4934304067698423,
"brier_soft": 0.429266777378387,
"nll": 0.8859090277806523,
"nll_soft": 0.9286131105577126,
"ece": 0.07366032000041674,
"mean_confidence": 0.6295792434553326,
"confident_error_rate": 0.001034126163391934,
"score_mae": 0.53153959018479
}
},
"by_source": {
"ag_news": {
"count": 100,
"accuracy": 0.88,
"brier": 0.1956816155108337,
"brier_soft": 0.1956816155108337,
"nll": 0.3569359667701418,
"nll_soft": 0.3569359667701418,
"ece": 0.09543950200080867,
"mean_confidence": 0.957247873544693,
"confident_error_rate": 0.06
},
"amazon_polarity": {
"count": 100,
"accuracy": 0.97,
"brier": 0.06277382037310442,
"brier_soft": 0.06277382037310442,
"nll": 0.14048191098950258,
"nll_soft": 0.14048191098950258,
"ece": 0.028436743021011374,
"mean_confidence": 0.9850584268569946,
"confident_error_rate": 0.03
},
"amazon_stars": {
"count": 100,
"accuracy": 0.53,
"brier": 0.576941525754432,
"brier_soft": 0.576941525754432,
"nll": 0.9323376081861305,
"nll_soft": 0.9323376081861305,
"ece": 0.16485882312059405,
"mean_confidence": 0.6662529197335243,
"confident_error_rate": 0.0,
"score_mae": 0.5123922033898998
},
"anli": {
"count": 100,
"accuracy": 0.39,
"brier": 1.0697171006877613,
"brier_soft": 1.0697171006877613,
"nll": 2.53446586810947,
"nll_soft": 2.53446586810947,
"ece": 0.49495991080999374,
"mean_confidence": 0.8732111909985543,
"confident_error_rate": 0.35
},
"arc_challenge": {
"count": 100,
"accuracy": 0.36,
"brier": 0.85564326196725,
"brier_soft": 0.85564326196725,
"nll": 1.6615319409245108,
"nll_soft": 1.6615319409245108,
"ece": 0.30022275894880296,
"mean_confidence": 0.6384962823987007,
"confident_error_rate": 0.07
},
"arc_easy": {
"count": 100,
"accuracy": 0.45,
"brier": 0.6727499751604388,
"brier_soft": 0.6727499751604388,
"nll": 1.3332344578928694,
"nll_soft": 1.3332344578928694,
"ece": 0.20594446569681168,
"mean_confidence": 0.6355124202370643,
"confident_error_rate": 0.03
},
"banking77": {
"count": 100,
"accuracy": 0.81,
"brier": 0.27520138323214327,
"brier_soft": 0.27520138323214327,
"nll": 0.7042584741411863,
"nll_soft": 0.7042584741411863,
"ece": 0.07309853732585905,
"mean_confidence": 0.7600076282024384,
"confident_error_rate": 0.0
},
"boolq": {
"count": 125,
"accuracy": 0.568,
"brier": 0.7086733078745265,
"brier_soft": 0.7086733078745265,
"nll": 1.0996114747593868,
"nll_soft": 1.0996114747593868,
"ece": 0.345228253364563,
"mean_confidence": 0.8863685808181763,
"confident_error_rate": 0.264,
"evidence_auc": 0.9706666666666667
},
"cb": {
"count": 28,
"accuracy": 0.9285714285714286,
"brier": 0.14187151621545094,
"brier_soft": 0.14187151621545094,
"nll": 0.3411688505063019,
"nll_soft": 0.3411688505063019,
"ece": 0.08071247168949669,
"mean_confidence": 0.9737525028841836,
"confident_error_rate": 0.07142857142857142
},
"civil_comments_severity": {
"count": 100,
"accuracy": 0.81,
"brier": 0.2641616792613485,
"brier_soft": 0.18398093610713928,
"nll": 0.48525818979939045,
"nll_soft": 0.6278173365191989,
"ece": 0.0930167320370674,
"mean_confidence": 0.7847322449088097,
"confident_error_rate": 0.01,
"score_mae": 0.31169346647828206
},
"civil_comments_toxic": {
"count": 100,
"accuracy": 0.94,
"brier": 0.08438325557871443,
"brier_soft": 0.02902029657244265,
"nll": 0.151836350238126,
"nll_soft": 0.27572009664481245,
"ece": 0.04216815173625941,
"mean_confidence": 0.9121545308828354,
"confident_error_rate": 0.01
},
"clinc150": {
"count": 100,
"accuracy": 0.8,
"brier": 0.2993320627901788,
"brier_soft": 0.2993320627901788,
"nll": 0.9652885590344843,
"nll_soft": 0.9652885590344843,
"ece": 0.11481654163450003,
"mean_confidence": 0.7029852807894349,
"confident_error_rate": 0.0
},
"cola": {
"count": 100,
"accuracy": 0.58,
"brier": 0.49509596184792515,
"brier_soft": 0.49509596184792515,
"nll": 0.6953869891066888,
"nll_soft": 0.6953869891066888,
"ece": 0.15781798183918,
"mean_confidence": 0.7369759422540665,
"confident_error_rate": 0.02
},
"commonsense_qa": {
"count": 100,
"accuracy": 0.42,
"brier": 0.7442126993153202,
"brier_soft": 0.7442126993153202,
"nll": 1.4986388927152847,
"nll_soft": 1.4986388927152847,
"ece": 0.2365888097882271,
"mean_confidence": 0.6122629001736641,
"confident_error_rate": 0.05
},
"cosmos_qa": {
"count": 120,
"accuracy": 0.4,
"brier": 0.7216093164068191,
"brier_soft": 0.7216093164068191,
"nll": 1.3523367352044129,
"nll_soft": 1.3523367352044129,
"ece": 0.22780664563179015,
"mean_confidence": 0.6138915047049522,
"confident_error_rate": 0.016666666666666666,
"evidence_auc": 0.836875
},
"dates": {
"count": 300,
"accuracy": 0.7633333333333333,
"brier": 0.28494306618151344,
"brier_soft": 0.28494306618151344,
"nll": 0.461646918030587,
"nll_soft": 0.461646918030587,
"ece": 0.06982577522595727,
"mean_confidence": 0.8165360381205876,
"confident_error_rate": 0.02666666666666667,
"score_mae": 0.42511637235032534
},
"dbpedia": {
"count": 100,
"accuracy": 0.99,
"brier": 0.01313320294332829,
"brier_soft": 0.01313320294332829,
"nll": 0.022723471569511666,
"nll_soft": 0.022723471569511666,
"ece": 0.015564617514610263,
"mean_confidence": 0.9883377414941787,
"confident_error_rate": 0.0
},
"dynasent": {
"count": 100,
"accuracy": 0.53,
"brier": 0.5607406074995769,
"brier_soft": 0.42609746333683257,
"nll": 0.9629181593054416,
"nll_soft": 1.0097704657215225,
"ece": 0.1584402248263359,
"mean_confidence": 0.6659586235880852,
"confident_error_rate": 0.0,
"score_mae": 0.5633883766271174
},
"enron_spam": {
"count": 100,
"accuracy": 0.94,
"brier": 0.08979354473427502,
"brier_soft": 0.08979354473427502,
"nll": 0.1666279776671364,
"nll_soft": 0.1666279776671364,
"ece": 0.0407087934017182,
"mean_confidence": 0.9807087934017181,
"confident_error_rate": 0.03
},
"go_emotions": {
"count": 100,
"accuracy": 0.54,
"brier": 0.5986230077830135,
"brier_soft": 0.23341383083194883,
"nll": 1.5085460014590555,
"nll_soft": 2.014061282637175,
"ece": 0.1133653888106346,
"mean_confidence": 0.47643122762441636,
"confident_error_rate": 0.0
},
"hellaswag": {
"count": 100,
"accuracy": 0.49,
"brier": 0.7303985853764171,
"brier_soft": 0.7303985853764171,
"nll": 1.4383446168093819,
"nll_soft": 1.4383446168093819,
"ece": 0.22660601407289505,
"mean_confidence": 0.658343857228756,
"confident_error_rate": 0.06
},
"imdb": {
"count": 100,
"accuracy": 0.95,
"brier": 0.08479113838840116,
"brier_soft": 0.08479113838840116,
"nll": 0.20038125969079196,
"nll_soft": 0.20038125969079196,
"ece": 0.04454047501087191,
"mean_confidence": 0.977543198466301,
"confident_error_rate": 0.03
},
"ledgar": {
"count": 100,
"accuracy": 0.75,
"brier": 0.37746571387860506,
"brier_soft": 0.37746571387860506,
"nll": 1.117831587098228,
"nll_soft": 1.117831587098228,
"ece": 0.09851389944553374,
"mean_confidence": 0.7810949206352233,
"confident_error_rate": 0.01
},
"liar": {
"count": 100,
"accuracy": 0.32,
"brier": 0.7427717464748098,
"brier_soft": 0.7427717464748098,
"nll": 1.4772028411140192,
"nll_soft": 1.4772028411140192,
"ece": 0.10479136735200882,
"mean_confidence": 0.3535214939713478,
"confident_error_rate": 0.0,
"score_mae": 0.9304733183607459
},
"measuring_hate_speech": {
"count": 100,
"accuracy": 0.83,
"brier": 0.26393481691390774,
"brier_soft": 0.1380680952422536,
"nll": 0.5018035055388043,
"nll_soft": 0.6680358527776807,
"ece": 0.1381094029545784,
"mean_confidence": 0.7157146391272545,
"confident_error_rate": 0.0,
"score_mae": 0.5079832007945515
},
"medqa": {
"count": 100,
"accuracy": 0.34,
"brier": 0.8253335318993641,
"brier_soft": 0.8253335318993641,
"nll": 1.6004826115846618,
"nll_soft": 1.6004826115846618,
"ece": 0.24258146941661834,
"mean_confidence": 0.5722464674711227,
"confident_error_rate": 0.01
},
"mnli": {
"count": 100,
"accuracy": 0.78,
"brier": 0.35049440453320635,
"brier_soft": 0.35049440453320635,
"nll": 0.7340184397638638,
"nll_soft": 0.7340184397638638,
"ece": 0.1504778018593788,
"mean_confidence": 0.9077262398600578,
"confident_error_rate": 0.09
},
"mrpc": {
"count": 100,
"accuracy": 0.81,
"brier": 0.296402084274162,
"brier_soft": 0.296402084274162,
"nll": 0.49396724593180946,
"nll_soft": 0.49396724593180946,
"ece": 0.12271335303783416,
"mean_confidence": 0.913284073472023,
"confident_error_rate": 0.11
},
"multirc": {
"count": 125,
"accuracy": 0.664,
"brier": 0.5778966045169998,
"brier_soft": 0.5778966045169998,
"nll": 1.0105245482832101,
"nll_soft": 1.0105245482832101,
"ece": 0.2395370769500732,
"mean_confidence": 0.8928435230255127,
"confident_error_rate": 0.232,
"evidence_auc": 0.6994666666666667
},
"openbookqa": {
"count": 100,
"accuracy": 0.52,
"brier": 0.6241696926704345,
"brier_soft": 0.6241696926704345,
"nll": 1.2342317482843357,
"nll_soft": 1.2342317482843357,
"ece": 0.1368167817592621,
"mean_confidence": 0.6306584364175797,
"confident_error_rate": 0.01
},
"piqa": {
"count": 100,
"accuracy": 0.61,
"brier": 0.5360872271281376,
"brier_soft": 0.5360872271281376,
"nll": 0.7988801723216932,
"nll_soft": 0.7988801723216932,
"ece": 0.1281885802745819,
"mean_confidence": 0.7082396709918976,
"confident_error_rate": 0.07
},
"policy_application": {
"count": 150,
"accuracy": 0.7333333333333333,
"brier": 0.3540402117052364,
"brier_soft": 0.3540402117052364,
"nll": 0.5500808292524187,
"nll_soft": 0.5500808292524187,
"ece": 0.07159115393956501,
"mean_confidence": 0.7941183785597483,
"confident_error_rate": 0.02666666666666667,
"evidence_auc": 0.5696533682145193
},
"policy_hard": {
"count": 300,
"accuracy": 0.6966666666666667,
"brier": 0.39680403372516454,
"brier_soft": 0.39680403372516454,
"nll": 0.6353135836137924,
"nll_soft": 0.6353135836137924,
"ece": 0.09296291559934615,
"mean_confidence": 0.7867085169752439,
"confident_error_rate": 0.03,
"evidence_auc": 0.7445023148148148
},
"qqp": {
"count": 100,
"accuracy": 0.8,
"brier": 0.33138385122291225,
"brier_soft": 0.33138385122291225,
"nll": 0.5997304360971085,
"nll_soft": 0.5997304360971085,
"ece": 0.160344073176384,
"mean_confidence": 0.92669453561306,
"confident_error_rate": 0.11
},
"race": {
"count": 125,
"accuracy": 0.464,
"brier": 0.6912585134522173,
"brier_soft": 0.6912585134522173,
"nll": 1.3744160623714958,
"nll_soft": 1.3744160623714958,
"ece": 0.14605404186248777,
"mean_confidence": 0.5852895030975341,
"confident_error_rate": 0.016,
"evidence_auc": 0.7856
},
"rotten_tomatoes": {
"count": 100,
"accuracy": 0.89,
"brier": 0.2377622469757273,
"brier_soft": 0.2377622469757273,
"nll": 0.3884358803375364,
"nll_soft": 0.3884358803375364,
"ece": 0.13390481829643247,
"mean_confidence": 0.7656551349163055,
"confident_error_rate": 0.0
},
"rte": {
"count": 100,
"accuracy": 0.7,
"brier": 0.48653410178986045,
"brier_soft": 0.48653410178986045,
"nll": 0.8552582691879445,
"nll_soft": 0.8552582691879445,
"ece": 0.20795070528984066,
"mean_confidence": 0.9079507052898407,
"confident_error_rate": 0.2
},
"scotus": {
"count": 100,
"accuracy": 0.63,
"brier": 0.558399670895838,
"brier_soft": 0.558399670895838,
"nll": 1.3936418468284624,
"nll_soft": 1.3936418468284624,
"ece": 0.13719820365309715,
"mean_confidence": 0.6636926798522472,
"confident_error_rate": 0.01
},
"shp": {
"count": 100,
"accuracy": 0.5,
"brier": 0.6054587301920193,
"brier_soft": 0.2182656459890131,
"nll": 0.8379870996959774,
"nll_soft": 0.8549041361552857,
"ece": 0.22597848415374755,
"mean_confidence": 0.7121235334873199,
"confident_error_rate": 0.04
},
"sms_spam": {
"count": 100,
"accuracy": 0.96,
"brier": 0.1060751114641901,
"brier_soft": 0.1060751114641901,
"nll": 0.21970222462417918,
"nll_soft": 0.21970222462417918,
"ece": 0.1075866764783859,
"mean_confidence": 0.852413323521614,
"confident_error_rate": 0.0
},
"snli": {
"count": 100,
"accuracy": 0.79,
"brier": 0.3950873837225307,
"brier_soft": 0.3950873837225307,
"nll": 0.8923272739656278,
"nll_soft": 0.8923272739656278,
"ece": 0.20375718295574188,
"mean_confidence": 0.9457461875677109,
"confident_error_rate": 0.16
},
"social_iqa": {
"count": 100,
"accuracy": 0.5,
"brier": 0.7282720596366673,
"brier_soft": 0.7282720596366673,
"nll": 1.4898040089024713,
"nll_soft": 1.4898040089024713,
"ece": 0.2804051974415779,
"mean_confidence": 0.766021781861782,
"confident_error_rate": 0.11
},
"sst2": {
"count": 100,
"accuracy": 0.9,
"brier": 0.22684034833869313,
"brier_soft": 0.22684034833869313,
"nll": 0.3794984833755713,
"nll_soft": 0.3794984833755713,
"ece": 0.14272936522960666,
"mean_confidence": 0.7780753713846207,
"confident_error_rate": 0.02
},
"sst5": {
"count": 100,
"accuracy": 0.56,
"brier": 0.5610144765744216,
"brier_soft": 0.5610144765744216,
"nll": 1.012664727296375,
"nll_soft": 1.012664727296375,
"ece": 0.06806655168533324,
"mean_confidence": 0.5649158990383149,
"confident_error_rate": 0.0,
"score_mae": 0.5462690752366325
},
"stance_abortion": {
"count": 60,
"accuracy": 0.6333333333333333,
"brier": 0.524088736443351,
"brier_soft": 0.524088736443351,
"nll": 0.8961191172069207,
"nll_soft": 0.8961191172069207,
"ece": 0.21964709063371024,
"mean_confidence": 0.8473132938146591,
"confident_error_rate": 0.11666666666666667
},
"stance_atheism": {
"count": 60,
"accuracy": 0.7,
"brier": 0.4295024180903305,
"brier_soft": 0.4295024180903305,
"nll": 0.7695298592508046,
"nll_soft": 0.7695298592508046,
"ece": 0.20804943144321447,
"mean_confidence": 0.8740352998177211,
"confident_error_rate": 0.11666666666666667
},
"stance_feminist": {
"count": 60,
"accuracy": 0.6833333333333333,
"brier": 0.5069301963867938,
"brier_soft": 0.5069301963867938,
"nll": 0.8636041819841961,
"nll_soft": 0.8636041819841961,
"ece": 0.24766051123539606,
"mean_confidence": 0.8220602666338285,
"confident_error_rate": 0.08333333333333333
},
"stance_hillary": {
"count": 60,
"accuracy": 0.6666666666666666,
"brier": 0.5074511953888173,
"brier_soft": 0.5074511953888173,
"nll": 0.9144219988342865,
"nll_soft": 0.9144219988342865,
"ece": 0.1895925045013428,
"mean_confidence": 0.8227309505144755,
"confident_error_rate": 0.08333333333333333
},
"stsb": {
"count": 100,
"accuracy": 0.45,
"brier": 0.6911273352227522,
"brier_soft": 0.4113556479959866,
"nll": 1.3398778933038693,
"nll_soft": 1.3971825733832768,
"ece": 0.17724637597799303,
"mean_confidence": 0.6206034556031227,
"confident_error_rate": 0.0,
"score_mae": 0.6759577402091236
},
"subjectivity": {
"count": 100,
"accuracy": 0.94,
"brier": 0.0767058129522083,
"brier_soft": 0.0767058129522083,
"nll": 0.11295796720663813,
"nll_soft": 0.11295796720663813,
"ece": 0.04773887455463412,
"mean_confidence": 0.9554775649309158,
"confident_error_rate": 0.0
},
"swag": {
"count": 100,
"accuracy": 0.4,
"brier": 0.8060758700239877,
"brier_soft": 0.8060758700239877,
"nll": 1.6377803414833798,
"nll_soft": 1.6377803414833798,
"ece": 0.313101019859314,
"mean_confidence": 0.713101019859314,
"confident_error_rate": 0.08
},
"trec": {
"count": 100,
"accuracy": 0.96,
"brier": 0.0627841559130499,
"brier_soft": 0.0627841559130499,
"nll": 0.1128754384695211,
"nll_soft": 0.1128754384695211,
"ece": 0.03479043036699291,
"mean_confidence": 0.9710938653349876,
"confident_error_rate": 0.01
},
"tweet_hate": {
"count": 100,
"accuracy": 0.54,
"brier": 0.4972376556033321,
"brier_soft": 0.4972376556033321,
"nll": 0.692663662352403,
"nll_soft": 0.692663662352403,
"ece": 0.11306986272335055,
"mean_confidence": 0.6452960926294327,
"confident_error_rate": 0.0
},
"tweet_sentiment": {
"count": 100,
"accuracy": 0.67,
"brier": 0.4101129435493609,
"brier_soft": 0.4101129435493609,
"nll": 0.6681253383561087,
"nll_soft": 0.6681253383561087,
"ece": 0.06605960577726364,
"mean_confidence": 0.6303303089737892,
"confident_error_rate": 0.0,
"score_mae": 0.3951122248638421
},
"vitaminc": {
"count": 100,
"accuracy": 0.55,
"brier": 0.6667891043228811,
"brier_soft": 0.6667891043228811,
"nll": 1.3148888047347953,
"nll_soft": 1.3148888047347953,
"ece": 0.2980202031135559,
"mean_confidence": 0.8368437862396241,
"confident_error_rate": 0.16
},
"wic": {
"count": 100,
"accuracy": 0.61,
"brier": 0.5528101823846308,
"brier_soft": 0.5528101823846308,
"nll": 0.8525569252877085,
"nll_soft": 0.8525569252877085,
"ece": 0.21445172727108003,
"mean_confidence": 0.8244517272710801,
"confident_error_rate": 0.12
},
"winogrande": {
"count": 100,
"accuracy": 0.57,
"brier": 0.5353339564884243,
"brier_soft": 0.5353339564884243,
"nll": 0.7633381528774766,
"nll_soft": 0.7633381528774766,
"ece": 0.17423450052738187,
"mean_confidence": 0.7043132489919662,
"confident_error_rate": 0.04
},
"yelp": {
"count": 100,
"accuracy": 0.62,
"brier": 0.41939405623046616,
"brier_soft": 0.41939405623046616,
"nll": 0.7083493947055206,
"nll_soft": 0.7083493947055206,
"ece": 0.12632228136062626,
"mean_confidence": 0.6600254368782044,
"confident_error_rate": 0.0,
"score_mae": 0.4118902616520063
}
},
"reliability": [
{
"bin": "0.0-0.1",
"count": 4,
"confidence": 0.06522660795599222,
"accuracy": 0.5
},
{
"bin": "0.1-0.2",
"count": 11,
"confidence": 0.16455867480147968,
"accuracy": 0.18181818181818182
},
{
"bin": "0.2-0.3",
"count": 88,
"confidence": 0.2596369771794839,
"accuracy": 0.29545454545454547
},
{
"bin": "0.3-0.4",
"count": 172,
"confidence": 0.3557821914207104,
"accuracy": 0.25
},
{
"bin": "0.4-0.5",
"count": 412,
"confidence": 0.45668168980809093,
"accuracy": 0.40048543689320387
},
{
"bin": "0.5-0.6",
"count": 828,
"confidence": 0.5498056534839713,
"accuracy": 0.4782608695652174
},
{
"bin": "0.6-0.7",
"count": 719,
"confidence": 0.6487497972513605,
"accuracy": 0.5465924895688457
},
{
"bin": "0.7-0.8",
"count": 714,
"confidence": 0.7503752866212059,
"accuracy": 0.6162464985994398
},
{
"bin": "0.8-0.9",
"count": 827,
"confidence": 0.8522957843044793,
"accuracy": 0.727932285368803
},
{
"bin": "0.9-1.0",
"count": 2338,
"confidence": 0.9691763384613449,
"accuracy": 0.8614200171086399
}
]
}