Tool-use Planning on ToolBench G2-Instruction
87.59Win RateGPT4 TOPGUN
Evaluation results
| Method | Links | |
|---|---|---|
| GPT4 TOPGUNReference=ChatGPT REACT2024.02 | 87.59 | |
| GPT4 TOPGUNReference=T.LLAMA REACT2024.02 | 86.24 | |
| GPT4 TOPGUNReference=T.LLAMA DFSDT+Ret2024.02 | 83.07 | |
| GPT4 TOPGUNReference=ChatGPT DFSDT2024.02 | 81.63 | |
| GPT4 TOPGUNReference=GPT4 ReACT2024.02 | 78.61 | |
| GPT4 TOPGUNReference=T.LLAMA DFSDT2024.02 | 78.31 | |
| GPT4 TOPGUNReference=GPT4 DFSDT2024.02 | 73.92 | |
| GPT4 DFSDTReference=ChatGPT REACT2024.02 | 73.3 | |
| ChatGPT DFSDTReference=ChatGPT REACT2024.02 | 72 | |
| T.LLAMA DFSDTReference=ChatGPT REACT2024.02 | 68.5 | |
| T.LLAMA DFSDT+RetReference=ChatGPT REACT2024.02 | 68.5 | |
| GPT4 ReACTReference=ChatGPT REACT2024.02 | 65.8 | |
| T.LLAMA REACTReference=ChatGPT REACT2024.02 | 50.8 |