AI learns to adapt reward preferences on the fly without retraining

Original: In-Context Reward Adaptation for Robust Preference Modeling

Source: arxiv.org ↗