MMLU
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
MMLU Pro
91.32CPT (95%)
4
MMLU-Redux
86.14CPT Score (95%)
4
MMLU
86.42CPT Accuracy (95%)
4
MMLU-Pro
0.786Hypervolume
4
MMLU-Redux
0.907Hypervolume
4
MMLU
86.09Hypervolume
4
MMLU OOD
52.58CPT Score (80%)
4
MMLU-Redux (test)
0.923Hypervolume
4
MMLU (test)
87.2Hypervolume
4
MMLU
67.6MMLU Score (%)
4
MMLU-Pro
40Accuracy
4
MMLU
86Accuracy
4
MMLU Out of Domain
1,896.9MTL Score
4
MMLU Math&Logic (train)
39.2R-PSR
4
MMLU (test)
95.77Accuracy
4
MMLU Target Subjects
52.19mEM
4
MMLU Redux
0.46Training Cost (USD)
4
MMLU
88.3Exact Match
4
MMLU
71.82Accuracy
4
MMLU Redux
67.3Accuracy
4
MMLU
58.4MMLU Knowledge Score
4
MMLU-Pro
84.43Accuracy
4
MMLU English
89.9Score
4
MMLU
58.16Accuracy
4
MMLU
4.3Normalized PLL Score
4