Visually Grounded Self-Reflection for Vision-Language Models via Reinforcement Learning figure
AlphaXiv 中文概览(可滚动查看)