Function Calling on BFCL (Held-In)
89.4AccuracySHAD+RFT
Evaluation Results
| Method | Links | |
|---|---|---|
| SHAD+RFTModel=LLaMA3.1-8B2024.12 | 89.4 | |
| RewardFTModel=LLaMA3-8B2024.12 | 89.3 | |
| SFTModel=LLaMA3.1-8B2024.12 | 89.3 | |
| RewardFTModel=LLaMA3.1-8B2024.12 | 88.2 | |
| SHAD+α-FTModel=LLaMA3.1-8B2024.12 | 88.2 | |
| SHAD+RFTModel=LLaMA3-8B2024.12 | 87.6 | |
| SHAD+α-FTModel=LLaMA3-8B2024.12 | 87.2 | |
| SFTModel=LLaMA3-8B2024.12 | 85.9 | |
| Rho-1Model=LLaMA3.1-8B2024.12 | 84.6 | |
| Regex+RFTModel=LLaMA3-8B2024.12 | 83.81 | |
| Rho-1Model=LLaMA3-8B2024.12 | 82.9 | |
| RegexModel=LLaMA3.1-8B2024.12 | 82.1 | |
| RegexModel=LLaMA3-8B2024.12 | 81 | |
| Regex+RFTModel=LLaMA3.1-8B2024.12 | 80.31 |