Safety Evaluation on Stereotype (test)
100Stereotype ScoreSFT
Evaluation Results
| Method | Links | |
|---|---|---|
| SFTBase Model=Qwen2.5-7B-Instruct, Alignment=SFT2026.02 | 100 | |
| SFT + OGPSABase Model=Qwen2.5-7B-Instruct, Alignment=SFT, Strategy=OGPSA2026.02 | 100 | |
| SFT-DPO + OGPSABase Model=Qwen2.5-7B-Instruct, Alignment=SFT-DPO, Strategy=OGPSA2026.02 | 100 | |
| SFT-DPO + General DataBase Model=Llama3.1-8B-Instruct, Alignment=SFT-DPO, Strategy=General Data2026.02 | 94.44 | |
| SFT + OGPSABase Model=Llama3.1-8B-Instruct, Alignment=SFT, Strategy=OGPSA2026.02 | 92.53 |