Exploring How Generative MLLMs Perceive More Than CLIP with the Same Vision Encoder.

Siting Li, Pang Wei Koh, Simon Shaolei Du

Browse the full ACL paper archive.