Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

KMMLU

Benchmarks

Task NameDataset NameSOTA ResultTrend
KnowledgeKMMLU
Knowledge EM77.9
9
Text Question AnsweringKMMLU-Redux
Accuracy51.8
8
Text Question AnsweringKMMLU-Pro
Accuracy46.85
8
KnowledgeKMMLU-Pro
Score70.9
7
General Knowledge EvaluationKMMLU
Accuracy55.23
5
Korean Language UnderstandingKMMLU-Pro
Accuracy73
5
Text-to-TextKMMLU Korean
Score75.2
4
General Language UnderstandingKMMLU-Pro
Score64
4
General Language UnderstandingKMMLU
Overall Score73
4
KnowledgeKMMLU Redux
Score75.9
3
KnowledgeKMMLU
Score0.787
3
Showing 11 of 11 rows