Visual Representation Alignment for Multimodal Large Language Models Paper • 2509.07979 • Published Sep 9 • 82