EpiPersona: Persona Projection and Episode Coupling for Pluralistic Preference Modeling

2026-03-30

AI Summary

Existing approaches struggle to disentangle users’ stable personality traits from their dynamic contextual preferences, limiting cross-scenario generalization. This work proposes a personality-context coupling framework that, for the first time, explicitly separates long-term personality characteristics from short-term contextual factors without requiring predefined preference dimensions. The method achieves this through low-dimensional personality projections and discrete code aggregation, and further introduces a context-personality coupling mechanism to enable context-aware preference modeling. Empirical results demonstrate that the proposed approach significantly outperforms current baselines, exhibiting particularly strong generalization performance under conditions of abrupt contextual shifts and data sparsity.

Abstract

Pluralistic alignment is essential for adapting large language models (LLMs) to the diverse preferences of individuals and minority groups. However, existing approaches often mix stable personal traits with episode-specific factors, limiting their ability to generalize across episodes. To address this challenge, we introduce EpiPersona, a framework for explicit persona-episode coupling. EpiPersona first projects noisy preference feedback into a low-dimensional persona space, where similar personas are aggregated into shared discrete codes. This process separates enduring personal characteristics from situational signals without relying on predefined preference dimensions. The inferred persona representation is then coupled with the current episode, enabling episode-aware preference prediction. Extensive experiments show that EpiPersona consistently outperforms the baselines. It achieves notable performance gains in hard episodic-shift scenarios, while remaining effective with sparse preference data.

Problem

• pluralistic preference modeling
• persona-episode coupling
• preference generalization
• episodic shift
• personal traits

Innovation

• persona projection
• episode coupling
• pluralistic preference modeling
• discrete persona codes
• preference disentanglement