Reasoning
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Reasoning Dataset
86.9Accuracy (Acc)
21
7 reasoning datasets
65.74Reasoning Accuracy
15
Reasoning
100Math Score
12
Reasoning BBH, GSM8k, MATH, Sciriff
79.65Accuracy
8
Reasoning
74.23ROUGE-1
8
Reasoning
194.21Throughput
8
Reasoning
72.2Average Score
5
Reasoning
0.76Spearman's Rho
5
Reasoning
3.51Bits per Token
5
Reasoning
0.6Attack Success Rate (ASR)
4