See Further, Think Deeper: Advancing VLM's Reasoning Ability with Low-level Visual Cues and Reflection

arXiv:2604.24339v1 Announce Type: cross Abstract: Recent advances in Vision-Language Models (VLMs) have benefited from Reinforcement Learning (RL) for enhanced reasoning. However, existing methods still face critical limitations, including the lack of low-level visual information and effective...