MTEB-lite: named reference models (models.csv), per-task calibration (calibration.csv), eng (v2) + eng-super (v3 selection with 46 reference models)
ac198e3 verified | task,task_type,main_score,estimator,K,a,b,n_fit_models,fit_pearson_r,fit_residual_sd,fitted_on | |
| ArguAna,Retrieval,ndcg_at_10,weighted_mean,100,-0.10005997732620146,1.0813815334607768,28,0.9860753855515598,0.01865286685807321,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| ArXivHierarchicalClusteringP2P,Clustering,v_measure,set_metric,100,-0.2976961904516839,1.2000771386547835,28,0.9754036802564997,0.00898049719012277,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| ArXivHierarchicalClusteringS2S,Clustering,v_measure,set_metric,100,-0.11323257760785495,0.9392913038278978,28,0.9711996270782994,0.00958570393928921,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| AskUbuntuDupQuestions,Reranking,map,weighted_mean,100,-0.017846758489783208,1.0272479409898907,27,0.9947854363408681,0.004745193868619753,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| BIOSSES,STS,cosine_spearman,set_metric,100,0.004938372988820848,0.9941255097207276,27,0.9879353524831916,0.00767000328060124,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| Banking77Classification,Classification,accuracy,weighted_mean,100,0.0014965336774522332,1.0086785725869465,27,0.9927016521875103,0.012107010396477498,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| BiorxivClusteringP2P.v2,Clustering,v_measure,set_metric,100,-0.17821729066134545,1.0822442930965996,28,0.9016496288896529,0.025039460208964436,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| CQADupstackGamingRetrieval,Retrieval,ndcg_at_10,weighted_mean,100,-0.23310134497588392,1.2066623727072237,21,0.8799109823901703,0.018119177291805736,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| CQADupstackUnixRetrieval,Retrieval,ndcg_at_10,weighted_mean,100,-0.019325104095223627,0.7734360729179047,21,0.9441767323555531,0.016435402684867973,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| ClimateFEVERHardNegatives,Retrieval,ndcg_at_10,weighted_mean,100,-0.2604054602467844,1.1610040350890722,22,0.953620606373082,0.018128907737792542,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| FEVERHardNegatives,Retrieval,ndcg_at_10,weighted_mean,100,-0.6707289732843521,1.7082563510742166,22,0.9931176591255516,0.00798972890130148,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| FiQA2018,Retrieval,ndcg_at_10,weighted_mean,100,-0.1369816183704981,0.9408666148590501,27,0.9907154783455474,0.014818619936612796,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| HotpotQAHardNegatives,Retrieval,ndcg_at_10,weighted_mean,100,-0.2678160144873314,1.2139809781665385,28,0.9965805948635259,0.009924199738408487,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| ImdbClassification,Classification,accuracy,weighted_mean,100,-0.029395046156113453,1.036078676998864,27,0.9916028548602781,0.014731955031563326,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| MTOPDomainClassification,Classification,accuracy,weighted_mean,100,0.021271246770427108,0.9736354910656642,27,0.9981709050312381,0.0057148522622465486,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| MassiveIntentClassification,Classification,accuracy,weighted_mean,100,0.021825334997885804,0.9639637349818443,28,0.9978270158124792,0.00815945024090585,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| MassiveScenarioClassification,Classification,accuracy,weighted_mean,100,0.007015225497673798,0.9774543822977197,28,0.9974800838245731,0.007451653419621154,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| MedrxivClusteringP2P.v2,Clustering,v_measure,set_metric,100,-0.598842863122897,1.6229157093578612,28,0.9235909157746447,0.015645273813104706,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| MedrxivClusteringS2S.v2,Clustering,v_measure,set_metric,100,-0.241689883432464,1.0652880110645917,28,0.9614321766766849,0.011333172488397516,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| MindSmallReranking,Reranking,map,weighted_mean,100,0.049427304970633634,0.8431011128330212,27,0.8786252198427205,0.006515692551166455,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| SCIDOCS,Retrieval,ndcg_at_10,weighted_mean,100,-0.011776045875905972,0.767525154960734,21,0.9591621214058388,0.00815712588831035,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| SICK-R,STS,cosine_spearman,set_metric,100,0.48783479500161564,0.6052777277801114,28,0.9163832773361791,0.019285363159615765,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| STS12,STS,cosine_spearman,set_metric,100,0.48955729144807164,0.5575881983517633,28,0.9775335336714752,0.013038719935479453,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| STS13,STS,cosine_spearman,set_metric,100,0.33649196261082537,0.6752101243696883,28,0.994942962493681,0.011555454385652576,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| STS14,STS,cosine_spearman,set_metric,100,0.3607938406745327,0.6488820075471318,28,0.9809931863688226,0.019861065279317604,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| STS15,STS,cosine_spearman,set_metric,100,0.4509161849821474,0.5992897929677832,28,0.9724680087623602,0.012622623548015936,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| STSBenchmark,STS,cosine_spearman,set_metric,100,0.4204671325482033,0.6158906454887774,28,0.976250584583606,0.013964771348771162,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| SprintDuplicateQuestions,PairClassification,max_ap,set_metric,37,0.778790435827451,0.28755256564350207,28,0.5962859661256674,0.08374206537708272,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| StackExchangeClustering.v2,Clustering,v_measure,set_metric,100,-7.688125512943829,9.045950170837848,28,0.6867200486049896,0.08197552150602591,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| StackExchangeClusteringP2P.v2,Clustering,v_measure,set_metric,100,-10.772405208562688,12.33572978873516,27,0.727502609287048,0.03801792913719614,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| TRECCOVID,Retrieval,ndcg_at_10,weighted_mean,50,-1.5161092582406308,2.577774315605494,22,0.9428690279617006,0.040805890163984775,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| Touche2020Retrieval.v3,Retrieval,ndcg_at_10,weighted_mean,49,-1.2787728950278578,2.3531429434591304,20,0.9203131547716173,0.021418646902364648,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| ToxicConversationsClassification,Classification,accuracy,weighted_mean,100,0.020319503164656156,0.9642828929928275,28,0.9855884794822495,0.014771368754537805,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| TweetSentimentExtractionClassification,Classification,accuracy,weighted_mean,100,-0.042099709342135874,1.0604488957136515,27,0.9932463625455074,0.009406941844391725,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| TwentyNewsgroupsClustering.v2,Clustering,v_measure,set_metric,100,-0.27998527506904747,1.240324750327259,27,0.8288502782890271,0.04975400445641078,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| TwitterSemEval2015,PairClassification,max_ap,set_metric,100,0.5429012003278054,0.439862379197883,27,0.4557074005416845,0.06683638895284126,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| TwitterURLCorpus,PairClassification,max_ap,set_metric,100,0.6986312189534242,0.47656090767545656,28,0.7499595862026545,0.017118857090771704,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| SummEvalSummarization.v2,Summarization,cosine_spearman,weighted_mean,100,0.008870940476184535,0.973266868083606,28,0.9981047114195146,0.0018425867121919142,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| AmazonCounterfactualClassification,Classification,accuracy,weighted_mean,100,-0.1131866105568904,1.1217291944802672,28,0.9958480478690829,0.00936810195724685,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| STS17,STS,cosine_spearman,set_metric,100,0.47745443922662756,0.548257173822908,28,0.9712432732273248,0.010979377740413524,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |
| STS22.v2,STS,cosine_spearman,set_metric,100,0.39558564411520386,0.6778900691097075,22,0.8162096447003683,0.012950219596595033,models.csv rows with role_eng = selection+calibration+LOMO-held-out (those with a full score for this task) | |