Mathematical Reasoning on DROP Numerical Answers Only (test)
72.3Accuracy (Exact Match)MetaMath
Evaluation Results
| Method | Links | |
|---|---|---|
| MetaMath#Params=70B, Evaluation Mode=Zero-shot2023.09 | 72.3 | |
| WizardMath#Params=70B, Evaluation Mode=Zero-shot2023.09 | 63.1 | |
| MetaMath#Params=13B, Evaluation Mode=Zero-shot2023.09 | 49.5 | |
| WizardMath#Params=13B, Evaluation Mode=Zero-shot2023.09 | 46.4 | |
| MetaMath#Params=7B, Evaluation Mode=Zero-shot2023.09 | 37.1 | |
| WizardMath#Params=7B, Evaluation Mode=Zero-shot2023.09 | 31.5 | |
| RFT#Params=7B, Evaluation Mode=Zero-shot2023.09 | 26.7 | |
| SFT#Params=7B, Evaluation Mode=Zero-shot2023.09 | 25.8 |