Chain-of-Spot: Interactive Reasoning Improves Large Vision-Language Models - Citation Graph | Papersgraph