Multimodal Preference Data Synthetic Alignment with Reward Model

MLLM
SyDa
Re-assign community