Mathematical Reasoning on KMath (pass@1)
75Pass@1GPT-4
Evaluation results
| Method | Links | |
|---|---|---|
| GPT-42023.10 | 75 | |
| KwaiYiiMath-HPA#Params=13B, Human preference alignment=true2023.10 | 74.47 | |
| KwaiYiiMath#Params=13B, Human preference alignment=false2023.10 | 73.4 | |
| Ernie Bot2023.10 | 72.87 | |
| ChatGPT2023.10 | 59.57 | |
| ChatGLM2#Params=6B2023.10 | 50 | |
| QWen#Params=7B2023.10 | 44.15 | |
| BaiChuan1#Params=13B2023.10 | 28.19 | |
| WizardMath#Params=13B2023.10 | 23.4 |