Reward Modeling and Preference Learning

Papers on training reward models and critics to learn human preferences through reinforcement learning, including approaches for visual reasoning, multi-objective alignment, and preference-based optimization.