List your company

Personalization on PRISM

81.62Personalization Win RateMIPO
5.315225.125144.93564.7449Mar 10, 2026

Evaluation results

MethodLinks
MIPOBase Model=Llama-3.2-3B-Instruct2026.0381.62
INFONCEBase Model=Llama-3.2-1B-Instruct2026.0380.93
MIPOBase Model=Llama-3.2-1B-Instruct2026.0380.93
RLAIF (MI)Base Model=Llama-3.2-3B-Instruct2026.0380.58
INFONCEBase Model=Llama-3.2-3B-Instruct2026.0380.58
RLAIFBase Model=Llama-3.2-3B-Instruct2026.0380.07
SFTBase Model=Llama-3.2-3B-Instruct2026.0378.52
SFT (Revision)Base Model=Llama-3.2-3B-Instruct2026.0378.35
Personalized PromptingBase Model=Llama-3.2-3B-Instruct2026.0376.8
SFT (Revision)Base Model=Llama-3.2-1B-Instruct2026.0373.37
PPO (MI)Base Model=Llama-3.2-1B-Instruct2026.0373.37
PPO (MI)Base Model=Llama-3.2-3B-Instruct2026.0373.02
Personalized PromptingBase Model=Llama-3.2-1B-Instruct2026.0372.17
SFTBase Model=Llama-3.2-1B-Instruct2026.0369.76
RLAIF (MI)Base Model=Llama-3.2-1B-Instruct2026.0369.07
RLAIF (MI)Base Model=Qwen2.5-7B-Instruct2026.0364.43
RLAIFBase Model=Qwen2.5-7B-Instruct2026.0363.4
MIPOBase Model=Qwen2.5-3B-Instruct2026.0359.79
SFT (Revision)Base Model=Qwen2.5-7B-Instruct2026.0359.28
MIPOBase Model=Qwen2.5-7B-Instruct2026.0358.76
Personalized PromptingBase Model=Qwen2.5-7B-Instruct2026.0358.25
INFONCEBase Model=Qwen2.5-7B-Instruct2026.0358.25
INFONCEBase Model=Qwen2.5-1.5B-Instruct2026.0357.87
RLAIFBase Model=Llama-3.2-1B-Instruct2026.0356.87
SFTBase Model=Qwen2.5-7B-Instruct2026.0356.7
MIPOBase Model=Qwen2.5-1.5B-Instruct2026.0356.53
PPO (MI)Base Model=Qwen2.5-7B-Instruct2026.0354.12
SFT (Revision)Base Model=Qwen2.5-3B-Instruct2026.0353.09
INFONCEBase Model=Qwen2.5-3B-Instruct2026.0351.89
RLAIF (MI)Base Model=Qwen2.5-3B-Instruct2026.0351.72
PPO (MI)Base Model=Qwen2.5-3B-Instruct2026.0350.52
Personalized PromptingBase Model=Qwen2.5-3B-Instruct2026.0349.49
RLAIFBase Model=Qwen2.5-3B-Instruct2026.0348.63
SFT (Revision)Base Model=Qwen2.5-1.5B-Instruct2026.0346.56
SFTBase Model=Qwen2.5-3B-Instruct2026.0344.67
Llama-3.2-3B-InstructBase Model=Llama-3.2-3B-Instruct2026.0340.72
Personalized PromptingBase Model=Qwen2.5-1.5B-Instruct2026.0339.18
Llama-3.2-1B-InstructBase Model=Llama-3.2-1B-Instruct2026.0336.6
SFTBase Model=Qwen2.5-1.5B-Instruct2026.0335.05
PPO (MI)Base Model=Qwen2.5-1.5B-Instruct2026.0332.47
Qwen2.5-7B-InstructBase Model=Qwen2.5-7B-Instruct2026.0320.62
RLAIF (MI)Base Model=Qwen2.5-1.5B-Instruct2026.0316.83
RLAIFBase Model=Qwen2.5-1.5B-Instruct2026.0313.92
Qwen2.5-3B-InstructBase Model=Qwen2.5-3B-Instruct2026.0311.86
Qwen2.5-1.5B-InstructBase Model=Qwen2.5-1.5B-Instruct2026.038.25