VisDoT : Enhancing Visual Reasoning through Human-Like Interpretation Grounding and Decomposition of Thought Paper • 2603.11631 • Published Mar 12 • 2
