Ground-R1: Incentivizing Grounded Visual Reasoning via Reinforcement Learning

Meng Cao, Haoze Zhao, Can Zhang +3

Large Vision-Language Models (LVLMs) have become powerful general-purpose assistants, yet their predictions often lack reliability and interpretability due to insufficient groundin…