GSM8K
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
GSM8K (test)
95.7Accuracy
11
GSM8K
10Error Rate
11
GSM8K
93.61Mean@4
11
GSM8K Instruct (test)
76.12Accuracy
11
GSM8K
79.68Accuracy
10
GSM8K
22.079Overall Accuracy
10
GSM8K
46.9EMA
10
GSM8K
4.28ECE
10
GSM8k (test)
94.84Accuracy
10
GSM8K
96.13Accuracy
10
GSM8K (test)
75.52Pass@1
10
GSM8K (test)
0.491Loss
10
GSM8K (test)
79Throughput (TPS)
10
GSM8k-Aug rationale format (original)
98.89Pass@1
10
GSM8K
100AR (%)
10
GSM8K
89.2Expected Utility
10
GSM8K
81.5Accuracy
10
GSM8K
75.2Accuracy
10
GSM8K (test)
4.598Test Perplexity
10
GSM8K
258.35Decoding Speed (tokens/s)
10
GSM8K
187.43Throughput (tokens/s)
10
GSM8K (test)
1BE Score
10
GSM8K (test)
7.95BE Score
10
GSM8K
99.6Topic Entropy
10
GSM8K
0False Positive Rate
10