Fetching the paper…

Vision-Language Models are Zero-Shot Reward Models for Reinforcement Learning · Around