Generative RLHF-V: Learning Principles from Multi-modal Human Preference

Training multi-modal large language models (MLLMs) that align with human intentions is a long-term challenge. Traditional score-only reward models for alignment suffer from low accuracy, weak generalization, and poor interpretability, blocking the progress of alignment methods, \textit{e. g. ,} reinforcement learning from human feedback (RLHF). Generative reward models (GRMs) leverage MLLMs' intrinsic reasoning capabilities to discriminate pair-wise responses, but their pair-wise paradigm makes it hard to generalize to learnable rewards. We introduce Generative RLHF-V, a novel alignment framework that integrates GRMs with multi-modal RLHF. We propose a two-stage pipeline: \textbf{multi-modal generative reward modeling from RL}, where RL guides GRMs to actively capture human intention, then predict the correct pair-wise scores; and \textbf{RL optimization from grouped comparison}, which enhances multi-modal RL scoring precision by grouped responses comparison. Experimental results demonstrate that, besides out-of-distribution generalization of RM discrimination, our framework improves 4 MLLMs' performance across 7 benchmarks by 18. 1\%, while the baseline RLHF is only 5. 3\%. We further validate that Generative RLHF-V achieves a near-linear improvement with an increasing number of candidate responses.

Authors

• Jiayi Zhou Institute for AI, Peking University State Key Laboratory of General Artificial Intelligence, Institute for AI, Peking University
• Jiaming Ji Institute for AI, Peking University State Key Laboratory of General Artificial Intelligence, Institute for AI, Peking University
• Boyuan Chen
• Jiapeng Sun
• wenqi chen
• Donghai Hong
• Sirui Han
• Yike Guo Hong Kong University of Science and Technology
• Yaodong Yang Institute for Artificial Intelligence, Peking University

Keywords

No keywords are indexed for this paper.