Loop to Look Better: Learning Adaptive Depth for Question-Guided Perception in Vision-Language Models

Wencheng Ye, Anning Hu, Hongye Fang, Tianyi Wang, Junchi Yan

LoopLook repeatedly refines visual features in response to a question, with a learned stopping rule and token-wise mixing of intermediate states. Adapts the vision encoder while keeping the language model frozen; the extra computation occurs before ordinary answer generation.

First public: October 22, 2026 · Added to this list: October 23, 2026