Tool-use Planning on ToolBench G2-Category
78.78Win RateGPT4 TOPGUN
Evaluation results
| Method | Links | |
|---|---|---|
| GPT4 TOPGUNReference=ChatGPT REACT2024.02 | 78.78 | |
| GPT4 TOPGUNReference=T.LLAMA REACT2024.02 | 77.71 | |
| GPT4 TOPGUNReference=GPT4 ReACT2024.02 | 73.75 | |
| GPT4 TOPGUNReference=ChatGPT DFSDT2024.02 | 73.07 | |
| GPT4 TOPGUNReference=T.LLAMA DFSDT+Ret2024.02 | 72.92 | |
| GPT4 TOPGUNReference=T.LLAMA DFSDT2024.02 | 71.8 | |
| GPT4 TOPGUNReference=GPT4 DFSDT2024.02 | 71.35 | |
| ChatGPT DFSDTReference=ChatGPT REACT2024.02 | 64.8 | |
| GPT4 DFSDTReference=ChatGPT REACT2024.02 | 63.3 | |
| T.LLAMA DFSDT+RetReference=ChatGPT REACT2024.02 | 60.8 | |
| GPT4 ReACTReference=ChatGPT REACT2024.02 | 60.3 | |
| T.LLAMA DFSDTReference=ChatGPT REACT2024.02 | 58 | |
| T.LLAMA REACTReference=ChatGPT REACT2024.02 | 41.8 |