listener-rewarded-vlm-preferences

【效率与协作】Aligns reasoning traces with final decisions in preference models using an independent frozen VLM as listener. Achieves 67.4% accuracy on ImageReward by enforcing consistency between explanations and choices. Use when training reward models for image generation quality where both reasoning quality and accuracy matter.

独立 GitHub 快照 · 自动分类 · 以原作者文档为准