SeeMe: Mitigating Hallucinations in Large Vision-Language Models through Effective Visual Token Engineering figure
AlphaXiv 中文概览(可滚动查看)