Re2A: Situated Conversational Recommendation via Rubric-based Preference Reasoning and Alignment
This work introduces rubric-based preference reasoning, which uses automated rubrics to guide the model toward producing explicit preference states, and proposes a preference-conditioned optimization to align response generation with dual objectives: user preference satisfaction and situation consistency.
Dongding Lin, Jian Wang, Xiao-Ya Zhao et al.
· 0 citations