Personalization on Community Alignment (CA)
93.67Personalization Win-RateMIPO
Evaluation Results
| Method | Links | |
|---|---|---|
| MIPOBase Model=Llama-3.2-1B-Instruct2026.03 | 93.67 | |
| MIPOBase Model=Llama-3.2-3B-Instruct2026.03 | 90.33 | |
| RLAIFBase Model=Llama-3.2-3B-Instruct2026.03 | 90.17 | |
| INFONCEBase Model=Llama-3.2-1B-Instruct2026.03 | 89.5 | |
| RLAIF (MI)Base Model=Qwen2.5-7B-Instruct2026.03 | 86 | |
| SFTBase Model=Llama-3.2-3B-Instruct2026.03 | 84.17 | |
| SFTBase Model=Llama-3.2-1B-Instruct2026.03 | 83.83 | |
| INFONCEBase Model=Llama-3.2-3B-Instruct2026.03 | 83.67 | |
| INFONCEBase Model=Qwen2.5-7B-Instruct2026.03 | 82.5 | |
| SFT (Revision)Base Model=Llama-3.2-3B-Instruct2026.03 | 81.67 | |
| MIPOBase Model=Qwen2.5-7B-Instruct2026.03 | 81.5 | |
| RLAIF (MI)Base Model=Llama-3.2-1B-Instruct2026.03 | 81 | |
| RLAIFBase Model=Qwen2.5-7B-Instruct2026.03 | 80.5 | |
| MIPOBase Model=Qwen2.5-3B-Instruct2026.03 | 80.33 | |
| PPO (MI)Base Model=Llama-3.2-1B-Instruct2026.03 | 80.17 | |
| SFT (Revision)Base Model=Llama-3.2-1B-Instruct2026.03 | 80 | |
| RLAIFBase Model=Llama-3.2-1B-Instruct2026.03 | 79.83 | |
| PPO (MI)Base Model=Qwen2.5-7B-Instruct2026.03 | 79 | |
| MIPOBase Model=Qwen2.5-1.5B-Instruct2026.03 | 78.83 | |
| Personalized PromptingBase Model=Qwen2.5-7B-Instruct2026.03 | 78.5 | |
| Personalized PromptingBase Model=Llama-3.2-1B-Instruct2026.03 | 78 | |
| Personalized PromptingBase Model=Llama-3.2-3B-Instruct2026.03 | 78 | |
| RLAIFBase Model=Qwen2.5-3B-Instruct2026.03 | 77.33 | |
| SFTBase Model=Qwen2.5-7B-Instruct2026.03 | 77 | |
| SFT (Revision)Base Model=Qwen2.5-7B-Instruct2026.03 | 77 | |
| Personalized PromptingBase Model=Qwen2.5-3B-Instruct2026.03 | 76 | |
| SFT (Revision)Base Model=Qwen2.5-3B-Instruct2026.03 | 74.83 | |
| RLAIF (MI)Base Model=Llama-3.2-3B-Instruct2026.03 | 74.5 | |
| RLAIF (MI)Base Model=Qwen2.5-3B-Instruct2026.03 | 74.5 | |
| SFTBase Model=Qwen2.5-3B-Instruct2026.03 | 74.33 | |
| PPO (MI)Base Model=Llama-3.2-3B-Instruct2026.03 | 72.67 | |
| PPO (MI)Base Model=Qwen2.5-3B-Instruct2026.03 | 72.67 | |
| INFONCEBase Model=Qwen2.5-3B-Instruct2026.03 | 70 | |
| SFT (Revision)Base Model=Qwen2.5-1.5B-Instruct2026.03 | 67 | |
| Personalized PromptingBase Model=Qwen2.5-1.5B-Instruct2026.03 | 63.5 | |
| SFTBase Model=Qwen2.5-1.5B-Instruct2026.03 | 60.5 | |
| RLAIF (MI)Base Model=Qwen2.5-1.5B-Instruct2026.03 | 59.83 | |
| INFONCEBase Model=Qwen2.5-1.5B-Instruct2026.03 | 58.83 | |
| RLAIFBase Model=Qwen2.5-1.5B-Instruct2026.03 | 50.33 | |
| PPO (MI)Base Model=Qwen2.5-1.5B-Instruct2026.03 | 39.67 | |
| Qwen2.5-7B-InstructBase Model=Qwen2.5-7B-Instruct2026.03 | 39 | |
| Llama-3.2-3B-InstructBase Model=Llama-3.2-3B-Instruct2026.03 | 37 | |
| Qwen2.5-3B-InstructBase Model=Qwen2.5-3B-Instruct2026.03 | 34 | |
| Llama-3.2-1B-InstructBase Model=Llama-3.2-1B-Instruct2026.03 | 31 | |
| Qwen2.5-1.5B-InstructBase Model=Qwen2.5-1.5B-Instruct2026.03 | 25.5 |