Fetching the paper…

Multimodal RewardBench: Holistic Evaluation of Reward Models for Vision Language Models · Around