MATH
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
MATH 500
97.3pass@1
8
MATH
43.66Accuracy
8
MATH500
61Accuracy
8
MATH level 1
79Accuracy
8
Math
60.8Math Score
8
MATH500
43.4Pass@1 Accuracy
8
Math
38.8BLEU
8
MATH
34Comparable SC (%)
8
MATH-500
66.91Pass@1
7
Math N=3.5k
85.2R@1
7
MATH500
0.89R^2 (Train)
7
MATH (forget)
0.3Accuracy
7
MATH 500
78.6Accuracy (MATH 500)
7
MATH 500
93.8Accuracy
7
Math
93.28Accuracy
7
MATH (test)
36.56Accuracy (MATH test)
7
MATH-500
87.17Accuracy (avg@K)
7
MATH 500
0.941Pass@32
7
MATH hard Level 4 5
76.3Accuracy (Exact Match)
7
MATH (test)
0.196Equation-level Diversity
7
MATH-500
94.35Accuracy
7
MATH
80Pass@1 Accuracy
7
Math datasets Aggregated models
0.058ECE2
7
MATH
76.1Accuracy
7
MATH500 0-shot (test)
40Pass@1 (Greedy)
7