Tool-augmented reasoning on MINT-Bench
9.85Success Rate (Turn 1)LLAMA PRO - INSTRUCT
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| LLAMA PRO - INSTRUCT2024.01 | 9.85 | 12.65 | 12.8 | 11.95 | 14.68 | 12.38 | |
| Mistral-Instruct-v0.12024.01 | 1.54 | 12.12 | 13.31 | 14.16 | 13.99 | 11.02 | |
| LLaMA2-7B-Chat2024.01 | 1.02 | 4.27 | 9.77 | 6.48 | 7.34 | 5.77 | |
| CodeLLaMA-7B-Instruct2024.01 | 0.34 | 7.85 | 10.24 | 9.73 | 8.7 | 7.37 | |
| AgentLM-7B2024.01 | 0 | 4.44 | 5.29 | 6.48 | 7.34 | 4.71 |