Mathematical Reasoning on GSM-SYS
80.9AccuracySATLM
Evaluation Results
| Method | Links | |
|---|---|---|
| SATLMLanguage Model=code-davinci-002, Decoding Method=self-consistency decoding2023.05 | 80.9 | |
| SATLMLanguage Model=code-davinci-002, Decoding Method=greedy decoding2023.05 | 69.4 | |
| COTLanguage Model=code-davinci-002, Decoding Method=self-consistency decoding2023.05 | 56.1 | |
| PROGLMLanguage Model=code-davinci-002, Decoding Method=self-consistency decoding2023.05 | 53.4 | |
| COTLanguage Model=code-davinci-002, Decoding Method=greedy decoding2023.05 | 46.5 | |
| PROGLMLanguage Model=code-davinci-002, Decoding Method=greedy decoding2023.05 | 43.4 | |
| STANDARDLanguage Model=code-davinci-002, Decoding Method=greedy decoding2023.05 | 21 |