MMLU
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
MMLU-Pro
49.4Science Accuracy
14
MMLU Pro
57.6Accuracy
14
MMLU Spanish (test)
77Accuracy
14
MMLU
64.29MMLU Score
14
MMLU (val)
69.12MMLU STEM Accuracy
14
MMLU
69.5MMLU Score
14
MMLU
61.9MMLU FL Score
14
MMLU-PRO (val)
91.5Top-1 Acc
14
MMLU-PRO
91.5Top-1 Acc
14
MMLU
58MMLU Factual Recall Accuracy
14
MMLU-Redux
83.7Accuracy
14
MMLU-Redux2
91.67PAQ
14
MMLU
65.98Vanilla Accuracy
14
MMLU STEM (test)
65.6Accuracy (ACC)
13
MMLU Social Sciences (test)
79.12Accuracy
13
MMLU Humanities (test)
62.47Accuracy (Humanities)
13
MMLU Five-client Non-IID (test)
48.5Law/Eth Score
13
MMLU-Redux
90.15Accuracy (English Knowledge)
13
MMLU STEM
89.1Accuracy
13
MMLU-Redux (test)
84.54Accuracy
13
MMLU
61.46MMLU Score
13
MMLU
89.2MMLU Accuracy
13
MMLU
75.5MMLU Accuracy
12
MMLU Pro
66.49Accuracy
12
MMLU-Pro
25.96Mean Score @16
12