GSM8K
Benchmarks
TaskDatasetSOTA resultTrendResults
GSM8K
5.51Speedup
61
GSM8K (test)
95.83Exact Match Accuracy (GSM8K Test)
60
GSM8K
93.63Exact Match Accuracy (GSM8K)
60
GSM8K (test)
96.23Accuracy
58
GSM8K
89.15Accuracy
57
GSM8K
98.8Accuracy (GSM8K)
55
GSM8K (test)
96.5Accuracy (GSM8K)
54
GSM8K
95.83Pass@4 Accuracy
54
GSM8K
0Student Leakage
54
GSM8K
95Pass@1 Accuracy
54
GSM8K official 1.3k set (test)
93.7Accuracy
53
GSM8k
3.79ASR@3
52
GSM8K 4-shot
86.13Score
51
GSM8K
96.7Accuracy
49
GSM8K (test)
79.01AUROC (Reference)
48
GSM8K
68.1Token Savings
48
GSM8K (test)
0.839Accuracy
48
GSM8K (test)
156.22Gain
47
GSM8K
96.51Pass@1
47
GSM8K 5-shot (test)
95.27Strict Match Accuracy
47
GSM8K
98.31Accuracy
46
GSM8K 5-shot
82.7Score
46
GSM8K
97.2Answer Accuracy
45
GSM8K (test)
232.39AUPC
44
GSM8K original (test)
95.2Accuracy
44