Safe RLHF-V: Safe Reinforcement Learning from Human Feedback in Multimodal Large Language Models

OffRL
Re-assign community