Utility Evaluation on Utility Benchmark
8.08Utility ScoreSFT
Evaluation Results
| Method | Links | |
|---|---|---|
| SFTModel=Qwen2.5-7B-Instruct2025.10 | 8.08 | |
| HarmRLVRModel=Qwen2.5-7B-Instruct2025.10 | 8.05 | |
| SFTModel=Qwen3-8B2025.10 | 8.02 | |
| BaseModel=Qwen2.5-7B-Instruct2025.10 | 7.98 | |
| BaseModel=Qwen3-8B2025.10 | 7.89 | |
| HarmRLVRModel=Qwen3-8B2025.10 | 7.81 | |
| SFTModel=Llama-3.1-8B-Instruct2025.10 | 7.79 | |
| HarmRLVRModel=Llama-3-8B-Instruct2025.10 | 7.73 | |
| BaseModel=Llama-3.1-8B-Instruct2025.10 | 7.73 | |
| HarmRLVRModel=Llama-3.1-8B-Instruct2025.10 | 7.71 | |
| BaseModel=Llama-3-8B-Instruct2025.10 | 7.48 | |
| SFTModel=Llama-3-8B-Instruct2025.10 | 7.44 | |
| HarmRLVRModel=DeepSeek-R1-Distill-Llama-8B2025.10 | 7.3 | |
| BaseModel=DeepSeek-R1-Distill-Llama-8B2025.10 | 7.22 | |
| SFTModel=DeepSeek-R1-Distill-Llama-8B2025.10 | 6.96 |