Fetching the paper…

Safe RLHF-V: Safe Reinforcement Learning from Multi-modal Human Feedback · Around