assay-compiled-base / eval-dev-scaled.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": 0.44247259100885217,
"brier_soft": 0.4170925201205985,
"nll": 0.8056102230141763,
"nll_soft": 0.8172796926231564,
"ece": 0.03653387851957047,
"mean_confidence": 0.6869317729442912,
"confident_error_rate": 0.019466710289546867,
"score_mae": 0.5729998939493892,
"evidence_auc": 0.9074727460810058
},
"by_type": {
"bool": {
"count": 1898,
"accuracy": 0.7787144362486829,
"brier": 0.3124533326117464,
"brier_soft": 0.30893223729451025,
"nll": 0.47617615577105815,
"nll_soft": 0.4805717663652272,
"ece": 0.04409051484755141,
"mean_confidence": 0.8090805982576155,
"confident_error_rate": 0.026343519494204427,
"evidence_auc": 0.9039322111530974
},
"choice": {
"count": 3248,
"accuracy": 0.625615763546798,
"brier": 0.5012455732095099,
"brier_soft": 0.4763148831509033,
"nll": 0.9690121828222874,
"nll_soft": 0.9798443434525502,
"ece": 0.06097491611004007,
"mean_confidence": 0.6575062932025669,
"confident_error_rate": 0.021243842364532018,
"evidence_auc": 0.9208606009429315
},
"score": {
"count": 967,
"accuracy": 0.5925542916235781,
"brier": 0.5002615323221618,
"brier_soft": 0.4304678889742543,
"nll": 0.9033726782062063,
"nll_soft": 0.9321314590592249,
"ece": 0.06700552232571198,
"mean_confidence": 0.546017075691376,
"confident_error_rate": 0.0,
"score_mae": 0.5729998939493892
}
},
"by_source": {
"ag_news": {
"count": 100,
"accuracy": 0.88,
"brier": 0.17829138819549176,
"brier_soft": 0.17829138819549176,
"nll": 0.30741759767501897,
"nll_soft": 0.30741759767501897,
"ece": 0.09543892771005633,
"mean_confidence": 0.9135349568724632,
"confident_error_rate": 0.02
},
"amazon_polarity": {
"count": 100,
"accuracy": 0.97,
"brier": 0.06162859960932998,
"brier_soft": 0.06162859960932998,
"nll": 0.13110608215243919,
"nll_soft": 0.13110608215243919,
"ece": 0.03008812248706818,
"mean_confidence": 0.9551744455099106,
"confident_error_rate": 0.03
},
"amazon_stars": {
"count": 100,
"accuracy": 0.53,
"brier": 0.5461979259182673,
"brier_soft": 0.5461979259182673,
"nll": 0.9164086895783538,
"nll_soft": 0.9164086895783538,
"ece": 0.11809196531772617,
"mean_confidence": 0.5753743493556976,
"confident_error_rate": 0.0,
"score_mae": 0.5362134778685868
},
"anli": {
"count": 100,
"accuracy": 0.39,
"brier": 0.9722760942391393,
"brier_soft": 0.9722760942391393,
"nll": 1.8542452552224933,
"nll_soft": 1.8542452552224933,
"ece": 0.44132104188203813,
"mean_confidence": 0.7956731787323952,
"confident_error_rate": 0.23
},
"arc_challenge": {
"count": 100,
"accuracy": 0.36,
"brier": 0.778481672987934,
"brier_soft": 0.778481672987934,
"nll": 1.4337150394138498,
"nll_soft": 1.4337150394138498,
"ece": 0.20447848409414293,
"mean_confidence": 0.5459534320235252,
"confident_error_rate": 0.0
},
"arc_easy": {
"count": 100,
"accuracy": 0.45,
"brier": 0.6407232552939703,
"brier_soft": 0.6407232552939703,
"nll": 1.2136572969992905,
"nll_soft": 1.2136572969992905,
"ece": 0.12803085684776305,
"mean_confidence": 0.5452363789081573,
"confident_error_rate": 0.01
},
"banking77": {
"count": 100,
"accuracy": 0.81,
"brier": 0.3814467708771144,
"brier_soft": 0.3814467708771144,
"nll": 1.0443211262890961,
"nll_soft": 1.0443211262890961,
"ece": 0.3126603952050208,
"mean_confidence": 0.4973396047949791,
"confident_error_rate": 0.0
},
"boolq": {
"count": 125,
"accuracy": 0.568,
"brier": 0.6168742406476001,
"brier_soft": 0.6168742406476001,
"nll": 0.8592622690829735,
"nll_soft": 0.8592622690829735,
"ece": 0.2938120722770691,
"mean_confidence": 0.8219961438179016,
"confident_error_rate": 0.032,
"evidence_auc": 0.9706666666666667
},
"cb": {
"count": 28,
"accuracy": 0.9285714285714286,
"brier": 0.13537618609721863,
"brier_soft": 0.13537618609721863,
"nll": 0.2816298666367622,
"nll_soft": 0.2816298666367622,
"ece": 0.11475402755396709,
"mean_confidence": 0.9356890469789505,
"confident_error_rate": 0.03571428571428571
},
"civil_comments_severity": {
"count": 100,
"accuracy": 0.81,
"brier": 0.28384330937032604,
"brier_soft": 0.17998165817214293,
"nll": 0.5338759677740593,
"nll_soft": 0.6298814779743555,
"ece": 0.14874677449464796,
"mean_confidence": 0.6996492806077004,
"confident_error_rate": 0.0,
"score_mae": 0.39768598369671965
},
"civil_comments_toxic": {
"count": 100,
"accuracy": 0.94,
"brier": 0.09920447825135499,
"brier_soft": 0.03237408913021391,
"nll": 0.19394092520219666,
"nll_soft": 0.2773696142795253,
"ece": 0.08483366131782531,
"mean_confidence": 0.8621529483795166,
"confident_error_rate": 0.0
},
"clinc150": {
"count": 100,
"accuracy": 0.8,
"brier": 0.4848727215372612,
"brier_soft": 0.4848727215372612,
"nll": 1.4845052623480186,
"nll_soft": 1.4845052623480186,
"ece": 0.4221790521405637,
"mean_confidence": 0.37782094785943626,
"confident_error_rate": 0.0
},
"cola": {
"count": 100,
"accuracy": 0.58,
"brier": 0.4671416305229769,
"brier_soft": 0.4671416305229769,
"nll": 0.6581302967113183,
"nll_soft": 0.6581302967113183,
"ece": 0.10910327613353732,
"mean_confidence": 0.6739702433347702,
"confident_error_rate": 0.0
},
"commonsense_qa": {
"count": 100,
"accuracy": 0.42,
"brier": 0.7026358240551468,
"brier_soft": 0.7026358240551468,
"nll": 1.3662261855444524,
"nll_soft": 1.3662261855444524,
"ece": 0.13439485043287275,
"mean_confidence": 0.510890220105648,
"confident_error_rate": 0.0
},
"cosmos_qa": {
"count": 120,
"accuracy": 0.4,
"brier": 0.674820119653572,
"brier_soft": 0.674820119653572,
"nll": 1.2437770795447451,
"nll_soft": 1.2437770795447451,
"ece": 0.14261294777194658,
"mean_confidence": 0.5219608165323735,
"confident_error_rate": 0.0,
"evidence_auc": 0.836875
},
"dates": {
"count": 300,
"accuracy": 0.7633333333333333,
"brier": 0.2819022563327028,
"brier_soft": 0.2819022563327028,
"nll": 0.4583224548625654,
"nll_soft": 0.4583224548625654,
"ece": 0.06106376767158511,
"mean_confidence": 0.7644337038199107,
"confident_error_rate": 0.0033333333333333335,
"score_mae": 0.45998074431250346
},
"dbpedia": {
"count": 100,
"accuracy": 0.99,
"brier": 0.015198128030094994,
"brier_soft": 0.015198128030094994,
"nll": 0.06724704860844338,
"nll_soft": 0.06724704860844338,
"ece": 0.057107390761375415,
"mean_confidence": 0.9441341024637222,
"confident_error_rate": 0.0
},
"dynasent": {
"count": 100,
"accuracy": 0.53,
"brier": 0.5435009917424021,
"brier_soft": 0.40578378452816594,
"nll": 0.9203439216635296,
"nll_soft": 0.95189629595177,
"ece": 0.12150101184844973,
"mean_confidence": 0.5809222316741943,
"confident_error_rate": 0.0,
"score_mae": 0.590812379643321
},
"enron_spam": {
"count": 100,
"accuracy": 0.94,
"brier": 0.08228255110990408,
"brier_soft": 0.08228255110990408,
"nll": 0.15493662205431416,
"nll_soft": 0.15493662205431416,
"ece": 0.043621286749839804,
"mean_confidence": 0.9467802709341049,
"confident_error_rate": 0.02
},
"go_emotions": {
"count": 100,
"accuracy": 0.54,
"brier": 0.6768268805616047,
"brier_soft": 0.25233950625259904,
"nll": 1.7380711253413985,
"nll_soft": 2.0785070377760406,
"ece": 0.24022679880261422,
"mean_confidence": 0.29977320119738576,
"confident_error_rate": 0.0
},
"hellaswag": {
"count": 100,
"accuracy": 0.49,
"brier": 0.6818457284598648,
"brier_soft": 0.6818457284598648,
"nll": 1.27200660086301,
"nll_soft": 1.27200660086301,
"ece": 0.1305046308040619,
"mean_confidence": 0.5641709685325622,
"confident_error_rate": 0.0
},
"imdb": {
"count": 100,
"accuracy": 0.95,
"brier": 0.08363715307563817,
"brier_soft": 0.08363715307563817,
"nll": 0.16551026463952032,
"nll_soft": 0.16551026463952032,
"ece": 0.026311837434768674,
"mean_confidence": 0.9574666857719422,
"confident_error_rate": 0.03
},
"ledgar": {
"count": 100,
"accuracy": 0.75,
"brier": 0.42560135728608983,
"brier_soft": 0.42560135728608983,
"nll": 1.3302157486233597,
"nll_soft": 1.3302157486233597,
"ece": 0.23709699064493178,
"mean_confidence": 0.5129030093550682,
"confident_error_rate": 0.0
},
"liar": {
"count": 100,
"accuracy": 0.32,
"brier": 0.7427052798517789,
"brier_soft": 0.7427052798517789,
"nll": 1.4964320047893989,
"nll_soft": 1.4964320047893989,
"ece": 0.055253805071115485,
"mean_confidence": 0.3017434249818325,
"confident_error_rate": 0.0,
"score_mae": 0.9733948946371674
},
"measuring_hate_speech": {
"count": 100,
"accuracy": 0.83,
"brier": 0.3177919283014086,
"brier_soft": 0.16750927245698513,
"nll": 0.5942829577379074,
"nll_soft": 0.7062310284569752,
"ece": 0.21628696441650394,
"mean_confidence": 0.6214796364307403,
"confident_error_rate": 0.0,
"score_mae": 0.6045454696938396
},
"medqa": {
"count": 100,
"accuracy": 0.34,
"brier": 0.7671574185751032,
"brier_soft": 0.7671574185751032,
"nll": 1.4356347250931116,
"nll_soft": 1.4356347250931116,
"ece": 0.17744469463825224,
"mean_confidence": 0.48414860367774964,
"confident_error_rate": 0.0
},
"mnli": {
"count": 100,
"accuracy": 0.78,
"brier": 0.3263025000138572,
"brier_soft": 0.3263025000138572,
"nll": 0.5982821287502831,
"nll_soft": 0.5982821287502831,
"ece": 0.10382685691118237,
"mean_confidence": 0.8509401044249535,
"confident_error_rate": 0.06
},
"mrpc": {
"count": 100,
"accuracy": 0.81,
"brier": 0.27720433744929446,
"brier_soft": 0.27720433744929446,
"nll": 0.43150783049960595,
"nll_soft": 0.43150783049960595,
"ece": 0.1067017036676407,
"mean_confidence": 0.856079712510109,
"confident_error_rate": 0.03
},
"multirc": {
"count": 125,
"accuracy": 0.664,
"brier": 0.5262665529453465,
"brier_soft": 0.5262665529453465,
"nll": 0.7910576977190109,
"nll_soft": 0.7910576977190109,
"ece": 0.20780441761016843,
"mean_confidence": 0.833528317451477,
"confident_error_rate": 0.136,
"evidence_auc": 0.6994666666666667
},
"openbookqa": {
"count": 100,
"accuracy": 0.52,
"brier": 0.6059996266277589,
"brier_soft": 0.6059996266277589,
"nll": 1.1493342555724386,
"nll_soft": 1.1493342555724386,
"ece": 0.08220899760723116,
"mean_confidence": 0.5423529654741287,
"confident_error_rate": 0.01
},
"piqa": {
"count": 100,
"accuracy": 0.61,
"brier": 0.5129767773118842,
"brier_soft": 0.5129767773118842,
"nll": 0.7296690239966398,
"nll_soft": 0.7296690239966398,
"ece": 0.14158266067504877,
"mean_confidence": 0.6558456552028656,
"confident_error_rate": 0.02
},
"policy_application": {
"count": 150,
"accuracy": 0.7333333333333333,
"brier": 0.3508971131550242,
"brier_soft": 0.3508971131550242,
"nll": 0.5516997947737176,
"nll_soft": 0.5516997947737176,
"ece": 0.044622519016265856,
"mean_confidence": 0.7177301915486654,
"confident_error_rate": 0.0,
"evidence_auc": 0.5696533682145193
},
"policy_hard": {
"count": 300,
"accuracy": 0.6966666666666667,
"brier": 0.38472149078986073,
"brier_soft": 0.38472149078986073,
"nll": 0.6128855879532513,
"nll_soft": 0.6128855879532513,
"ece": 0.053346475064754505,
"mean_confidence": 0.7120696005225181,
"confident_error_rate": 0.0,
"evidence_auc": 0.7445023148148148
},
"qqp": {
"count": 100,
"accuracy": 0.8,
"brier": 0.3034980053199748,
"brier_soft": 0.3034980053199748,
"nll": 0.4749638870383975,
"nll_soft": 0.4749638870383975,
"ece": 0.10882365942001343,
"mean_confidence": 0.8906454944610596,
"confident_error_rate": 0.07
},
"race": {
"count": 125,
"accuracy": 0.464,
"brier": 0.670541267736231,
"brier_soft": 0.670541267736231,
"nll": 1.2785957499648075,
"nll_soft": 1.2785957499648075,
"ece": 0.09217529821395876,
"mean_confidence": 0.4943129940032959,
"confident_error_rate": 0.0,
"evidence_auc": 0.7856
},
"rotten_tomatoes": {
"count": 100,
"accuracy": 0.89,
"brier": 0.26950262705978917,
"brier_soft": 0.26950262705978917,
"nll": 0.43769943345994206,
"nll_soft": 0.43769943345994206,
"ece": 0.18772910118103028,
"mean_confidence": 0.7022708988189698,
"confident_error_rate": 0.0
},
"rte": {
"count": 100,
"accuracy": 0.7,
"brier": 0.4417545164022784,
"brier_soft": 0.4417545164022784,
"nll": 0.6803569597273692,
"nll_soft": 0.6803569597273692,
"ece": 0.166305216550827,
"mean_confidence": 0.8475300312042237,
"confident_error_rate": 0.07
},
"scotus": {
"count": 100,
"accuracy": 0.63,
"brier": 0.57591555494716,
"brier_soft": 0.57591555494716,
"nll": 1.4268808895864786,
"nll_soft": 1.4268808895864786,
"ece": 0.1622882364690304,
"mean_confidence": 0.479786828905344,
"confident_error_rate": 0.0
},
"shp": {
"count": 100,
"accuracy": 0.5,
"brier": 0.553928533015661,
"brier_soft": 0.16866709422112394,
"nll": 0.7553610522259013,
"nll_soft": 0.7667537170621911,
"ece": 0.22228121280670163,
"mean_confidence": 0.6585486650466919,
"confident_error_rate": 0.0
},
"sms_spam": {
"count": 100,
"accuracy": 0.96,
"brier": 0.14873415443052604,
"brier_soft": 0.14873415443052604,
"nll": 0.29604541382027116,
"nll_soft": 0.29604541382027116,
"ece": 0.18373807787895202,
"mean_confidence": 0.7762619221210479,
"confident_error_rate": 0.0
},
"snli": {
"count": 100,
"accuracy": 0.79,
"brier": 0.3657109430484185,
"brier_soft": 0.3657109430484185,
"nll": 0.6824574908252256,
"nll_soft": 0.6824574908252256,
"ece": 0.1515702518820763,
"mean_confidence": 0.8926358130574227,
"confident_error_rate": 0.1
},
"social_iqa": {
"count": 100,
"accuracy": 0.5,
"brier": 0.6701157287052262,
"brier_soft": 0.6701157287052262,
"nll": 1.2110427461546294,
"nll_soft": 1.2110427461546294,
"ece": 0.21912074834108355,
"mean_confidence": 0.68725660353899,
"confident_error_rate": 0.04
},
"sst2": {
"count": 100,
"accuracy": 0.9,
"brier": 0.25762914103330403,
"brier_soft": 0.25762914103330403,
"nll": 0.42289569521733256,
"nll_soft": 0.42289569521733256,
"ece": 0.182864231467247,
"mean_confidence": 0.717135768532753,
"confident_error_rate": 0.0
},
"sst5": {
"count": 100,
"accuracy": 0.56,
"brier": 0.5750112386200766,
"brier_soft": 0.5750112386200766,
"nll": 1.0410791576793528,
"nll_soft": 1.0410791576793528,
"ece": 0.13715695679187778,
"mean_confidence": 0.4760879367589951,
"confident_error_rate": 0.0,
"score_mae": 0.569280381789431
},
"stance_abortion": {
"count": 60,
"accuracy": 0.6333333333333333,
"brier": 0.4775618932889926,
"brier_soft": 0.4775618932889926,
"nll": 0.772960340218129,
"nll_soft": 0.772960340218129,
"ece": 0.14922663221756619,
"mean_confidence": 0.7609471683700879,
"confident_error_rate": 0.05
},
"stance_atheism": {
"count": 60,
"accuracy": 0.7,
"brier": 0.39241280131421197,
"brier_soft": 0.39241280131421197,
"nll": 0.6474431235930713,
"nll_soft": 0.6474431235930713,
"ece": 0.15528117964665092,
"mean_confidence": 0.8136728132764498,
"confident_error_rate": 0.016666666666666666
},
"stance_feminist": {
"count": 60,
"accuracy": 0.6833333333333333,
"brier": 0.46696109115797474,
"brier_soft": 0.46696109115797474,
"nll": 0.7586194826479168,
"nll_soft": 0.7586194826479168,
"ece": 0.17778196881214778,
"mean_confidence": 0.7421373461683591,
"confident_error_rate": 0.05
},
"stance_hillary": {
"count": 60,
"accuracy": 0.6666666666666666,
"brier": 0.47286586681141635,
"brier_soft": 0.47286586681141635,
"nll": 0.8005868818543999,
"nll_soft": 0.8005868818543999,
"ece": 0.15976376285155613,
"mean_confidence": 0.7363345429301262,
"confident_error_rate": 0.03333333333333333
},
"stsb": {
"count": 100,
"accuracy": 0.45,
"brier": 0.6576776158521066,
"brier_soft": 0.3746345989346838,
"nll": 1.2299245867552293,
"nll_soft": 1.2685160423963167,
"ece": 0.10637196481227874,
"mean_confidence": 0.523349083662033,
"confident_error_rate": 0.0,
"score_mae": 0.7103589298564476
},
"subjectivity": {
"count": 100,
"accuracy": 0.94,
"brier": 0.07403801155263033,
"brier_soft": 0.07403801155263033,
"nll": 0.12967172016931594,
"nll_soft": 0.12967172016931594,
"ece": 0.06520502805709842,
"mean_confidence": 0.9230094337463379,
"confident_error_rate": 0.0
},
"swag": {
"count": 100,
"accuracy": 0.4,
"brier": 0.7323198866636677,
"brier_soft": 0.7323198866636677,
"nll": 1.3764612983642295,
"nll_soft": 1.3764612983642295,
"ece": 0.2521471258997917,
"mean_confidence": 0.6150306752324104,
"confident_error_rate": 0.02
},
"trec": {
"count": 100,
"accuracy": 0.96,
"brier": 0.06216060648768522,
"brier_soft": 0.06216060648768522,
"nll": 0.14523731763009665,
"nll_soft": 0.14523731763009665,
"ece": 0.07524359405040741,
"mean_confidence": 0.9198704010248184,
"confident_error_rate": 0.0
},
"tweet_hate": {
"count": 100,
"accuracy": 0.54,
"brier": 0.4864483510328587,
"brier_soft": 0.4864483510328587,
"nll": 0.6796653519155663,
"nll_soft": 0.6796653519155663,
"ece": 0.08831145048141475,
"mean_confidence": 0.6011694383621216,
"confident_error_rate": 0.0
},
"tweet_sentiment": {
"count": 100,
"accuracy": 0.67,
"brier": 0.4380984174109661,
"brier_soft": 0.4380984174109661,
"nll": 0.729035173618918,
"nll_soft": 0.729035173618918,
"ece": 0.13779532372951508,
"mean_confidence": 0.5552096617221832,
"confident_error_rate": 0.0,
"score_mae": 0.4073845294490457
},
"vitaminc": {
"count": 100,
"accuracy": 0.55,
"brier": 0.6059178299160775,
"brier_soft": 0.6059178299160775,
"nll": 1.0501780738072244,
"nll_soft": 1.0501780738072244,
"ece": 0.20950709015131,
"mean_confidence": 0.75950709015131,
"confident_error_rate": 0.07
},
"wic": {
"count": 100,
"accuracy": 0.61,
"brier": 0.502306829144201,
"brier_soft": 0.502306829144201,
"nll": 0.7209994732293087,
"nll_soft": 0.7209994732293087,
"ece": 0.16384762406349185,
"mean_confidence": 0.7613926827907562,
"confident_error_rate": 0.04
},
"winogrande": {
"count": 100,
"accuracy": 0.57,
"brier": 0.5063494004660651,
"brier_soft": 0.5063494004660651,
"nll": 0.7075272167886625,
"nll_soft": 0.7075272167886625,
"ece": 0.09909610986709594,
"mean_confidence": 0.6518526268005371,
"confident_error_rate": 0.0
},
"yelp": {
"count": 100,
"accuracy": 0.62,
"brier": 0.43856781301069964,
"brier_soft": 0.43856781301069964,
"nll": 0.7669727900329802,
"nll_soft": 0.7669727900329802,
"ece": 0.15332692295312883,
"mean_confidence": 0.5684746101498603,
"confident_error_rate": 0.0,
"score_mae": 0.44304582916665824
}
},
"reliability": [
{
"bin": "0.0-0.1",
"count": 17,
"confidence": 0.06687228787032996,
"accuracy": 0.35294117647058826
},
{
"bin": "0.1-0.2",
"count": 42,
"confidence": 0.16727064549922943,
"accuracy": 0.2619047619047619
},
{
"bin": "0.2-0.3",
"count": 183,
"confidence": 0.25834722043386577,
"accuracy": 0.3770491803278688
},
{
"bin": "0.3-0.4",
"count": 348,
"confidence": 0.3574332141841965,
"accuracy": 0.3994252873563218
},
{
"bin": "0.4-0.5",
"count": 677,
"confidence": 0.4517646449971586,
"accuracy": 0.4564254062038405
},
{
"bin": "0.5-0.6",
"count": 1026,
"confidence": 0.5484959817769235,
"accuracy": 0.5487329434697856
},
{
"bin": "0.6-0.7",
"count": 832,
"confidence": 0.6478659043518397,
"accuracy": 0.6334134615384616
},
{
"bin": "0.7-0.8",
"count": 756,
"confidence": 0.7482921844594693,
"accuracy": 0.7314814814814815
},
{
"bin": "0.8-0.9",
"count": 814,
"confidence": 0.8511582021338348,
"accuracy": 0.7457002457002457
},
{
"bin": "0.9-1.0",
"count": 1418,
"confidence": 0.9543087374844571,
"accuracy": 0.9160789844851904
}
]
}