VisOnlyQA: Large Vision Language Models Still Struggle with Visual Perception of Geometric Information Paper • 2412.00947 • Published 11 days ago • 7