List your company

Preference Labeling on Summarization

89Preference Labeling AccuracyCurriculum-RLAIF
66.1272.067883.94May 26, 2025

Evaluation results

MethodLinks
Curriculum-RLAIFBase Model=LLaMA-3-8B2025.0589
Conventional RLAIFBase Model=LLaMA-3-8B2025.0582
Curriculum-RLAIFBase Model=Gemma-1-2B2025.0579
RLCDBase Model=LLaMA-3-8B2025.0578
RLCDBase Model=Gemma-1-2B2025.0572
Conventional RLAIFBase Model=Gemma-1-2B2025.0571
CAIBase Model=LLaMA-3-8B2025.0570
CAIBase Model=Gemma-1-2B2025.0567