SemVink: Advancing VLMs' Semantic Understanding of Optical Illusions via Visual Global Thinking.

Sifan Li, Yujun Cai, Yiwei Wang

Browse the full EMNLP paper archive.