assay-1.7b / eval-dev-scaled.json
Berk's picture
Upload folder using huggingface_hub
7c1fb34 verified
Raw History Blame Contribute Delete
22.2 kB
{
"overall": {
"count": 5513,
"accuracy": 0.7395247596589879,
"brier": 0.35543221635694794,
"brier_soft": 0.32787097437269497,
"nll": 0.6798980190388866,
"nll_soft": 0.6939236417123891,
"ece": 0.03505060344509826,
"mean_confidence": 0.755017923742952,
"confident_error_rate": 0.02920370034463994,
"score_mae": 0.5833856387926468,
"evidence_auc": 0.9994043337529221
},
"by_type": {
"choice": {
"count": 2993,
"accuracy": 0.7213498162378884,
"brier": 0.3711201333461809,
"brier_soft": 0.3442827979081974,
"nll": 0.74394983123531,
"nll_soft": 0.7572803793083344,
"ece": 0.04781638815122127,
"mean_confidence": 0.7496169757884853,
"confident_error_rate": 0.024724356832609422,
"evidence_auc": 0.9993569870249167
},
"noul": {
"count": 1620,
"accuracy": 0.8345679012345679,
"brier": 0.253961842505383,
"brier_soft": 0.2492008712245667,
"nll": 0.44479597034730256,
"nll_soft": 0.44916685601652256,
"ece": 0.06477943753706905,
"mean_confidence": 0.8877689307040277,
"confident_error_rate": 0.053703703703703705,
"evidence_auc": 0.9991164126352993
},
"score": {
"count": 900,
"accuracy": 0.6288888888888889,
"brier": 0.485907849791127,
"brier_soft": 0.41489872910403874,
"nll": 0.8900738467905202,
"nll_soft": 0.9237895052708771,
"ece": 0.09814345782042143,
"mean_confidence": 0.5340272636882586,
"confident_error_rate": 0.0,
"score_mae": 0.5833856387926468
}
},
"by_source": {
"ag_news": {
"count": 100,
"accuracy": 0.88,
"brier": 0.1626150587313346,
"brier_soft": 0.1626150587313346,
"nll": 0.30714077176782956,
"nll_soft": 0.30714077176782956,
"ece": 0.06605552225768023,
"mean_confidence": 0.9010962930653971,
"confident_error_rate": 0.03
},
"amazon_polarity": {
"count": 100,
"accuracy": 0.94,
"brier": 0.0807016434249766,
"brier_soft": 0.0807016434249766,
"nll": 0.12384397712211918,
"nll_soft": 0.12384397712211918,
"ece": 0.035107052853219106,
"mean_confidence": 0.9569018952897135,
"confident_error_rate": 0.0
},
"amazon_stars": {
"count": 100,
"accuracy": 0.59,
"brier": 0.5273514854384125,
"brier_soft": 0.5273514854384125,
"nll": 0.9298323955017933,
"nll_soft": 0.9298323955017933,
"ece": 0.08849946708455993,
"mean_confidence": 0.525445663454732,
"confident_error_rate": 0.0,
"score_mae": 0.5409659532487517
},
"anli": {
"count": 100,
"accuracy": 0.48,
"brier": 0.7580110428459982,
"brier_soft": 0.7580110428459982,
"nll": 1.368622216757406,
"nll_soft": 1.368622216757406,
"ece": 0.3121258033428911,
"mean_confidence": 0.7723577728427032,
"confident_error_rate": 0.13
},
"arc_challenge": {
"count": 100,
"accuracy": 0.81,
"brier": 0.24055934761415848,
"brier_soft": 0.24055934761415848,
"nll": 0.4411706000362115,
"nll_soft": 0.4411706000362115,
"ece": 0.0673066479533461,
"mean_confidence": 0.8316750467569213,
"confident_error_rate": 0.0
},
"arc_easy": {
"count": 100,
"accuracy": 0.9,
"brier": 0.1375523622489258,
"brier_soft": 0.1375523622489258,
"nll": 0.2746098585419268,
"nll_soft": 0.2746098585419268,
"ece": 0.06593661835503903,
"mean_confidence": 0.8765803273420658,
"confident_error_rate": 0.0
},
"banking77": {
"count": 100,
"accuracy": 0.66,
"brier": 0.4431805276953609,
"brier_soft": 0.4431805276953609,
"nll": 1.1794235133838844,
"nll_soft": 1.1794235133838844,
"ece": 0.15041971992111725,
"mean_confidence": 0.7271155569329256,
"confident_error_rate": 0.02
},
"boolq": {
"count": 125,
"accuracy": 0.672,
"brier": 0.5943519230409933,
"brier_soft": 0.5943519230409933,
"nll": 1.198119047336582,
"nll_soft": 1.198119047336582,
"ece": 0.2934385165179548,
"mean_confidence": 0.9305267511946171,
"confident_error_rate": 0.264,
"evidence_auc": 1.0
},
"cb": {
"count": 28,
"accuracy": 0.9285714285714286,
"brier": 0.1206399809111701,
"brier_soft": 0.1206399809111701,
"nll": 0.21512117628452895,
"nll_soft": 0.21512117628452895,
"ece": 0.08924438988716607,
"mean_confidence": 0.9300469646070304,
"confident_error_rate": 0.0
},
"civil_comments_severity": {
"count": 100,
"accuracy": 0.85,
"brier": 0.2695113879269868,
"brier_soft": 0.1842131970836758,
"nll": 0.5509239546378076,
"nll_soft": 0.6722600300908189,
"ece": 0.14800686874253358,
"mean_confidence": 0.7138696562789161,
"confident_error_rate": 0.0,
"score_mae": 0.4315058978345452
},
"civil_comments_toxic": {
"count": 100,
"accuracy": 0.92,
"brier": 0.12684197636177458,
"brier_soft": 0.04971424161255073,
"nll": 0.2250471634876677,
"nll_soft": 0.29585551132903215,
"ece": 0.053707607791622175,
"mean_confidence": 0.8804237395801962,
"confident_error_rate": 0.01
},
"clinc150": {
"count": 100,
"accuracy": 0.72,
"brier": 0.40672308629755116,
"brier_soft": 0.40672308629755116,
"nll": 1.2759781850846823,
"nll_soft": 1.2759781850846823,
"ece": 0.15851017418590968,
"mean_confidence": 0.7813792962015765,
"confident_error_rate": 0.04
},
"cola": {
"count": 100,
"accuracy": 0.72,
"brier": 0.4117765664582892,
"brier_soft": 0.4117765664582892,
"nll": 0.6207637174912584,
"nll_soft": 0.6207637174912584,
"ece": 0.13335671444872274,
"mean_confidence": 0.8533567144487229,
"confident_error_rate": 0.08
},
"commonsense_qa": {
"count": 100,
"accuracy": 0.79,
"brier": 0.28843537622683724,
"brier_soft": 0.28843537622683724,
"nll": 0.5630187107405263,
"nll_soft": 0.5630187107405263,
"ece": 0.08747740471844836,
"mean_confidence": 0.7833733258289846,
"confident_error_rate": 0.02
},
"cosmos_qa": {
"count": 120,
"accuracy": 0.5916666666666667,
"brier": 0.5201670608791354,
"brier_soft": 0.5201670608791354,
"nll": 0.9967133911913827,
"nll_soft": 0.9967133911913827,
"ece": 0.15986319837836777,
"mean_confidence": 0.738942816266573,
"confident_error_rate": 0.041666666666666664,
"evidence_auc": 0.9925
},
"dbpedia": {
"count": 100,
"accuracy": 0.99,
"brier": 0.011042004099136306,
"brier_soft": 0.011042004099136306,
"nll": 0.02104353360930288,
"nll_soft": 0.02104353360930288,
"ece": 0.01389831021923994,
"mean_confidence": 0.990020302708288,
"confident_error_rate": 0.0
},
"dynasent": {
"count": 100,
"accuracy": 0.66,
"brier": 0.4904603061732375,
"brier_soft": 0.3492007570551342,
"nll": 0.8315245856719652,
"nll_soft": 0.8558785430767095,
"ece": 0.13365715855069826,
"mean_confidence": 0.5811434655880939,
"confident_error_rate": 0.0,
"score_mae": 0.5566348787742988
},
"enron_spam": {
"count": 100,
"accuracy": 0.97,
"brier": 0.044212729061323425,
"brier_soft": 0.044212729061323425,
"nll": 0.10096603728495956,
"nll_soft": 0.10096603728495956,
"ece": 0.01711993302888352,
"mean_confidence": 0.9753496362181928,
"confident_error_rate": 0.01
},
"go_emotions": {
"count": 100,
"accuracy": 0.57,
"brier": 0.680845077018312,
"brier_soft": 0.2609842220598735,
"nll": 1.7377780873686146,
"nll_soft": 2.1303361224920976,
"ece": 0.26184079332116555,
"mean_confidence": 0.3081592066788344,
"confident_error_rate": 0.0
},
"hellaswag": {
"count": 100,
"accuracy": 0.82,
"brier": 0.24091543463389853,
"brier_soft": 0.24091543463389853,
"nll": 0.4584441520294868,
"nll_soft": 0.4584441520294868,
"ece": 0.07152204188632028,
"mean_confidence": 0.7853499832313032,
"confident_error_rate": 0.0
},
"imdb": {
"count": 100,
"accuracy": 0.94,
"brier": 0.10389960094912438,
"brier_soft": 0.10389960094912438,
"nll": 0.18813249438725196,
"nll_soft": 0.18813249438725196,
"ece": 0.05724087175277302,
"mean_confidence": 0.9601319692633947,
"confident_error_rate": 0.04
},
"ledgar": {
"count": 100,
"accuracy": 0.72,
"brier": 0.38034738636938825,
"brier_soft": 0.38034738636938825,
"nll": 1.063681018918801,
"nll_soft": 1.063681018918801,
"ece": 0.07838524729470606,
"mean_confidence": 0.7610558636919699,
"confident_error_rate": 0.02
},
"liar": {
"count": 100,
"accuracy": 0.27,
"brier": 0.7719490896567759,
"brier_soft": 0.7719490896567759,
"nll": 1.5958625278397007,
"nll_soft": 1.5958625278397007,
"ece": 0.033802663573387265,
"mean_confidence": 0.26573383000473283,
"confident_error_rate": 0.0,
"score_mae": 1.1542656605030175
},
"measuring_hate_speech": {
"count": 100,
"accuracy": 0.82,
"brier": 0.32241765969838126,
"brier_soft": 0.18209662000522567,
"nll": 0.5957862814286871,
"nll_soft": 0.7160770828698846,
"ece": 0.19588341038588403,
"mean_confidence": 0.6241165896141161,
"confident_error_rate": 0.0,
"score_mae": 0.6257405208951771
},
"medqa": {
"count": 100,
"accuracy": 0.52,
"brier": 0.6204688217755439,
"brier_soft": 0.6204688217755439,
"nll": 1.1441714418928755,
"nll_soft": 1.1441714418928755,
"ece": 0.07335443810322065,
"mean_confidence": 0.5279374556156311,
"confident_error_rate": 0.0
},
"mnli": {
"count": 100,
"accuracy": 0.89,
"brier": 0.17481065065993526,
"brier_soft": 0.17481065065993526,
"nll": 0.33065511645217294,
"nll_soft": 0.33065511645217294,
"ece": 0.07649681566926217,
"mean_confidence": 0.8625791144915169,
"confident_error_rate": 0.03
},
"mrpc": {
"count": 100,
"accuracy": 0.82,
"brier": 0.2540667926292604,
"brier_soft": 0.2540667926292604,
"nll": 0.4153874415696912,
"nll_soft": 0.4153874415696912,
"ece": 0.09072470674487898,
"mean_confidence": 0.7960669137761736,
"confident_error_rate": 0.01
},
"multirc": {
"count": 125,
"accuracy": 0.728,
"brier": 0.4982938010410545,
"brier_soft": 0.4982938010410545,
"nll": 1.1173533355293808,
"nll_soft": 1.1173533355293808,
"ece": 0.2368767428889841,
"mean_confidence": 0.9123281362860616,
"confident_error_rate": 0.192,
"evidence_auc": 0.9938666666666667
},
"openbookqa": {
"count": 100,
"accuracy": 0.73,
"brier": 0.35918319957697675,
"brier_soft": 0.35918319957697675,
"nll": 0.7108537776596335,
"nll_soft": 0.7108537776596335,
"ece": 0.11907511376015169,
"mean_confidence": 0.7974778297066559,
"confident_error_rate": 0.06
},
"piqa": {
"count": 100,
"accuracy": 0.79,
"brier": 0.32680077893343484,
"brier_soft": 0.32680077893343484,
"nll": 0.5077245732572071,
"nll_soft": 0.5077245732572071,
"ece": 0.11010356068721007,
"mean_confidence": 0.7685585521255288,
"confident_error_rate": 0.02
},
"policy_application": {
"count": 150,
"accuracy": 0.8466666666666667,
"brier": 0.21779322319452216,
"brier_soft": 0.21779322319452216,
"nll": 0.3452863418296437,
"nll_soft": 0.3452863418296437,
"ece": 0.0801742745253821,
"mean_confidence": 0.8475822661354353,
"confident_error_rate": 0.006666666666666667,
"evidence_auc": 0.9908436886854153
},
"qqp": {
"count": 100,
"accuracy": 0.9,
"brier": 0.18394250218095784,
"brier_soft": 0.18394250218095784,
"nll": 0.28979379627290724,
"nll_soft": 0.28979379627290724,
"ece": 0.09559768684175766,
"mean_confidence": 0.8727692218380053,
"confident_error_rate": 0.01
},
"race": {
"count": 125,
"accuracy": 0.688,
"brier": 0.40703814587255865,
"brier_soft": 0.40703814587255865,
"nll": 0.7537291180537239,
"nll_soft": 0.7537291180537239,
"ece": 0.09619744773901716,
"mean_confidence": 0.7766490781661468,
"confident_error_rate": 0.016,
"evidence_auc": 0.9928
},
"rotten_tomatoes": {
"count": 100,
"accuracy": 0.9,
"brier": 0.14010738440191914,
"brier_soft": 0.14010738440191914,
"nll": 0.2416655146105705,
"nll_soft": 0.2416655146105705,
"ece": 0.04597459019218179,
"mean_confidence": 0.9104264144844768,
"confident_error_rate": 0.02
},
"rte": {
"count": 100,
"accuracy": 0.85,
"brier": 0.22111515141522584,
"brier_soft": 0.22111515141522584,
"nll": 0.3433614260676825,
"nll_soft": 0.3433614260676825,
"ece": 0.05934966559327243,
"mean_confidence": 0.9061174995419405,
"confident_error_rate": 0.04
},
"scotus": {
"count": 100,
"accuracy": 0.26,
"brier": 0.8915802571770692,
"brier_soft": 0.8915802571770692,
"nll": 2.3878836951913804,
"nll_soft": 2.3878836951913804,
"ece": 0.10914746015346483,
"mean_confidence": 0.16521583892580075,
"confident_error_rate": 0.0
},
"shp": {
"count": 100,
"accuracy": 0.5,
"brier": 0.49990982000791506,
"brier_soft": 0.1165292253075058,
"nll": 0.6957887808957649,
"nll_soft": 0.7022140495979027,
"ece": 0.09007377936084067,
"mean_confidence": 0.5508245098606296,
"confident_error_rate": 0.0
},
"sms_spam": {
"count": 100,
"accuracy": 0.99,
"brier": 0.01945800378248445,
"brier_soft": 0.01945800378248445,
"nll": 0.0451107545075836,
"nll_soft": 0.0451107545075836,
"ece": 0.0038497034110411155,
"mean_confidence": 0.9921453540295321,
"confident_error_rate": 0.01
},
"snli": {
"count": 100,
"accuracy": 0.85,
"brier": 0.199062955024896,
"brier_soft": 0.199062955024896,
"nll": 0.3503996991590813,
"nll_soft": 0.3503996991590813,
"ece": 0.0664930712350235,
"mean_confidence": 0.8970722048593263,
"confident_error_rate": 0.05
},
"social_iqa": {
"count": 100,
"accuracy": 0.71,
"brier": 0.431037408095384,
"brier_soft": 0.431037408095384,
"nll": 0.7204327278269901,
"nll_soft": 0.7204327278269901,
"ece": 0.106641996998918,
"mean_confidence": 0.7628192791146897,
"confident_error_rate": 0.03
},
"sst2": {
"count": 100,
"accuracy": 0.94,
"brier": 0.09230732647344755,
"brier_soft": 0.09230732647344755,
"nll": 0.17918581044030002,
"nll_soft": 0.17918581044030002,
"ece": 0.03225442688620987,
"mean_confidence": 0.9634628465849301,
"confident_error_rate": 0.04
},
"sst5": {
"count": 100,
"accuracy": 0.61,
"brier": 0.5360312392825825,
"brier_soft": 0.5360312392825825,
"nll": 0.9667997383108035,
"nll_soft": 0.9667997383108035,
"ece": 0.13009597538068704,
"mean_confidence": 0.479904024619313,
"confident_error_rate": 0.0,
"score_mae": 0.5185995248390466
},
"stance_abortion": {
"count": 60,
"accuracy": 0.65,
"brier": 0.41941041091876696,
"brier_soft": 0.41941041091876696,
"nll": 0.6945807290012106,
"nll_soft": 0.6945807290012106,
"ece": 0.06253681805122369,
"mean_confidence": 0.6790044674495604,
"confident_error_rate": 0.0
},
"stance_atheism": {
"count": 60,
"accuracy": 0.7333333333333333,
"brier": 0.33573074707097367,
"brier_soft": 0.33573074707097367,
"nll": 0.5493844906099609,
"nll_soft": 0.5493844906099609,
"ece": 0.11774155629528504,
"mean_confidence": 0.7629887792123973,
"confident_error_rate": 0.016666666666666666
},
"stance_feminist": {
"count": 60,
"accuracy": 0.65,
"brier": 0.42391371238817716,
"brier_soft": 0.42391371238817716,
"nll": 0.6858903532741335,
"nll_soft": 0.6858903532741335,
"ece": 0.09564019985317251,
"mean_confidence": 0.6729372426738106,
"confident_error_rate": 0.016666666666666666
},
"stance_hillary": {
"count": 60,
"accuracy": 0.7166666666666667,
"brier": 0.3950376026145419,
"brier_soft": 0.3950376026145419,
"nll": 0.6753047409123483,
"nll_soft": 0.6753047409123483,
"ece": 0.14666177006731274,
"mean_confidence": 0.6947910284891209,
"confident_error_rate": 0.016666666666666666
},
"stsb": {
"count": 100,
"accuracy": 0.53,
"brier": 0.562591826514361,
"brier_soft": 0.2903885199851368,
"nll": 1.0047635453769461,
"nll_soft": 1.042223637401206,
"ece": 0.05010116262109817,
"mean_confidence": 0.5006841341226713,
"confident_error_rate": 0.0,
"score_mae": 0.5846652815901485
},
"subjectivity": {
"count": 100,
"accuracy": 0.92,
"brier": 0.10356259237332947,
"brier_soft": 0.10356259237332947,
"nll": 0.1536756285071837,
"nll_soft": 0.1536756285071837,
"ece": 0.06302427253308573,
"mean_confidence": 0.9584267044094797,
"confident_error_rate": 0.02
},
"swag": {
"count": 100,
"accuracy": 0.71,
"brier": 0.39564301574865673,
"brier_soft": 0.39564301574865673,
"nll": 0.718116512775688,
"nll_soft": 0.718116512775688,
"ece": 0.1196927981046649,
"mean_confidence": 0.7007635426910775,
"confident_error_rate": 0.0
},
"trec": {
"count": 100,
"accuracy": 0.93,
"brier": 0.09869043173228237,
"brier_soft": 0.09869043173228237,
"nll": 0.2255870758700891,
"nll_soft": 0.2255870758700891,
"ece": 0.03541382371297605,
"mean_confidence": 0.92494501233904,
"confident_error_rate": 0.01
},
"tweet_hate": {
"count": 100,
"accuracy": 0.66,
"brier": 0.4774603068268896,
"brier_soft": 0.4774603068268896,
"nll": 0.6801771939569775,
"nll_soft": 0.6801771939569775,
"ece": 0.1890020966071993,
"mean_confidence": 0.7647629815710008,
"confident_error_rate": 0.02
},
"tweet_sentiment": {
"count": 100,
"accuracy": 0.71,
"brier": 0.4240819345210271,
"brier_soft": 0.4240819345210271,
"nll": 0.7214182225919487,
"nll_soft": 0.7214182225919487,
"ece": 0.14850151921708057,
"mean_confidence": 0.5748318141162528,
"confident_error_rate": 0.0,
"score_mae": 0.3819244602989468
},
"vitaminc": {
"count": 100,
"accuracy": 0.67,
"brier": 0.4246865636582119,
"brier_soft": 0.4246865636582119,
"nll": 0.7577260270925018,
"nll_soft": 0.7577260270925018,
"ece": 0.13463918498472002,
"mean_confidence": 0.7641609989462581,
"confident_error_rate": 0.05
},
"wic": {
"count": 100,
"accuracy": 0.66,
"brier": 0.41189935927685106,
"brier_soft": 0.41189935927685106,
"nll": 0.5931627190857616,
"nll_soft": 0.5931627190857616,
"ece": 0.08766499905395608,
"mean_confidence": 0.6542995718931661,
"confident_error_rate": 0.0
},
"winogrande": {
"count": 100,
"accuracy": 0.58,
"brier": 0.5746250741107299,
"brier_soft": 0.5746250741107299,
"nll": 0.8577718414331656,
"nll_soft": 0.8577718414331656,
"ece": 0.19468332651889314,
"mean_confidence": 0.7540031893020952,
"confident_error_rate": 0.11
},
"yelp": {
"count": 100,
"accuracy": 0.62,
"brier": 0.4687757189083782,
"brier_soft": 0.4687757189083782,
"nll": 0.81375336975503,
"nll_soft": 0.81375336975503,
"ece": 0.12454148770521407,
"mean_confidence": 0.5405161953954996,
"confident_error_rate": 0.0,
"score_mae": 0.45616857114989023
}
},
"reliability": [
{
"bin": "0.0-0.1",
"count": 2,
"confidence": 0.09036682874041202,
"accuracy": 0.5
},
{
"bin": "0.1-0.2",
"count": 122,
"confidence": 0.16504008000438247,
"accuracy": 0.30327868852459017
},
{
"bin": "0.2-0.3",
"count": 124,
"confidence": 0.2443757508869671,
"accuracy": 0.2903225806451613
},
{
"bin": "0.3-0.4",
"count": 191,
"confidence": 0.35652105981608595,
"accuracy": 0.418848167539267
},
{
"bin": "0.4-0.5",
"count": 568,
"confidence": 0.46077405748501,
"accuracy": 0.47183098591549294
},
{
"bin": "0.5-0.6",
"count": 538,
"confidence": 0.5489642809707137,
"accuracy": 0.5557620817843866
},
{
"bin": "0.6-0.7",
"count": 556,
"confidence": 0.6459295074019494,
"accuracy": 0.6528776978417267
},
{
"bin": "0.7-0.8",
"count": 457,
"confidence": 0.7521871119337753,
"accuracy": 0.7177242888402626
},
{
"bin": "0.8-0.9",
"count": 678,
"confidence": 0.8530844565623106,
"accuracy": 0.8097345132743363
},
{
"bin": "0.9-1.0",
"count": 2277,
"confidence": 0.9685375709309736,
"accuracy": 0.9292929292929293
}
],
"seconds": 31.36005187034607,
"model": "runs/assay-1.7b-v2",
"data": "data/v2/dev.jsonl"
}