How Does Personalized Memory Shape LLM Behavior? Benchmarking Rational Preference Utilization in Personalized Assistants

2026-01-23

AI Summary

This work addresses the problem of irrational personalization in large language models, where the incorporation of personalized memory often introduces irrelevant information, thereby distorting user intent understanding and degrading user experience. To tackle this issue, the study presents the first systematic characterization of the phenomenon and proposes RP-Reasoner, a novel approach that models memory utilization as a pragmatic reasoning process to enable selective integration of personalized information. Furthermore, the authors introduce RPEval, the first benchmark specifically designed for evaluating rational preference utilization, comprising a personalized intent inference dataset and a multi-granularity evaluation protocol. Experimental results demonstrate that RP-Reasoner significantly outperforms strong baselines and successfully rectifies 80% of irrational personalization cases in a large-scale commercial assistant.

Technology Category

• Knowledge Representation and Reasoning: Preferences
• Machine Learning: Learning Preferences or Rankings
• Natural Language Processing: (Large) Language Models

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Personalized, context-aware and across-device search
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Large language model (LLM)-powered assistants have recently integrated memory mechanisms that record user preferences, leading to more personalized and user-aligned responses. However, irrelevant personalized memories are often introduced into the context, interfering with the LLM's intent understanding. To comprehensively investigate the dual effects of personalization, we develop RPEval, a benchmark comprising a personalized intent reasoning dataset and a multi-granularity evaluation protocol. RPEval reveals the widespread phenomenon of irrational personalization in existing LLMs and, through error pattern analysis, illustrates its negative impact on user experience. Finally, we introduce RP-Reasoner, which treats memory utilization as a pragmatic reasoning process, enabling the selective integration of personalized information. Experimental results demonstrate that our method significantly outperforms carefully designed baselines on RPEval, and resolves 80% of the bad cases observed in a large-scale commercial personalized assistant, highlighting the potential of pragmatic reasoning to mitigate irrational personalization. Our benchmark is publicly available at https://github.com/XueyangFeng/RPEval.

Problem

• personalized memory
• irrational personalization
• intent understanding
• large language models
• user preferences

Innovation

• personalized memory
• rational preference utilization
• pragmatic reasoning
• LLM personalization
• benchmarking