提出新基准与方法，解决大模型记忆干扰用户意图的问题
How Does Personalized Memory Shape LLM Behavior? Benchmarking Rational Preference Utilization in Personalized Assistants
- 构建多粒度评估框架RPEval，分析个性化记忆对意图理解的影响
- 发现80%商业助手的错误由不合理记忆导致，影响用户体验
- 设计RP-Reasoner，通过实用推理选择性使用记忆，显著提升表现
基于大语言模型的智能助手已集成记忆机制以记录用户偏好，实现更个性化的响应。然而，无关记忆常被引入上下文，干扰模型对用户意图的理解。为全面探究个性化带来的双重影响，我们构建了RPEval基准，包含个性化意图推理数据集和多粒度评估协议。实验揭示现有大模型普遍存在非理性个性化现象，误差模式分析表明其严重损害用户体验。为此，我们提出RP-Reasoner，将记忆利用视为一种实用推理过程，实现个性化信息的选择性整合。实验结果表明，该方法在RPEval上显著优于精心设计的基线模型，并解决了大规模商业个性化助手80%的不良案例，展示了实用推理在缓解非理性个性化方面的潜力。基准代码已开源：https://github.com/XueyangFeng/RPEval。
原文摘要 · Abstract (English)
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.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。