ruMTEB (v1.1) results

#1
by BorisTM - opened
Task Score
GeoreviewClassification 21.2158
HeadlineClassification 23.4912
InappropriatenessClassification 51.4648
KinopoiskClassification 34.0133
MassiveIntentClassification 15.1560
MassiveScenarioClassification 19.1995
RuReviewsClassification 39.0039
RuSciBenchGRNTIClassification 5.8154
RuSciBenchOECDClassification 5.0244
GeoreviewClusteringP2P 19.6829
RuSciBenchGRNTIClusteringP2P 9.3429
RuSciBenchOECDClusteringP2P 9.2425
CEDRClassification 33.2678
SensitiveTopicsClassification 17.6758
TERRa 52.5268
MIRACLReranking 11.4410
RuBQReranking 25.0730
MIRACLRetrievalHardNegatives.v2 1.1190
RiaNewsRetrievalHardNegatives.v2 1.3040
RuBQRetrieval 0.4220
RUParaPhraserSTS 41.2728
STS22 −5.2536
RuSTSBenchmarkSTS 44.1540
Mean(Task) 20.6807
Mean(TaskType) 22.9293

Sign up or log in to comment