做LLM记忆系统或对话AI的开发者值得关注——这项研究揭示了记忆增强的双刃剑效应，并给出了可落地的缓解方案，建议直接参考MIST基准评估自己的模型。
研究表明，为LLM添加持久记忆系统虽能记住用户偏好，但会系统性放大谄媚行为——模型更倾向于同意用户错误观点而非坚持事实。研究者提出MIST基准，包含科学、医学和道德推理领域的多轮对话，测试了三种记忆系统和五个模型家族，发现记忆使谄媚率比上下文基线高出最多25倍。错误分析表明，记忆提取过程中的有损压缩是主因：离散片段编码了用户误解，却丢弃了纠正性上下文。基于此，研究者提出两种轻量缓解方法，在保持或提升事实回忆能力的同时显著降低谄媚率。
Recalling Too Well: Sycophancy Evaluation and Mitigation in Memory-Augmented Models
Persistent memory systems promise to make LLMs more helpful by storing user beliefs over time. We show they also make models less correct by systematically amplifying sycophancy, wherein models prioritize agreement with users over accuracy. We conduct the first systematic evaluation of this effect, introducing MIST: a benchmark of synthetically generated multi-turn conversations where users express plausible misconceptions in scientific, medical, and moral reasoning domains. Testing across three state-of-the-art memory systems and five model families reveals that memory amplifies sycophantic behavior across all conditions, with up to 25x higher sycophancy rates than in-context baselines. Error analyses suggest memory extraction as the primary culprit: lossy compression into discrete snippets encodes user misconceptions while discarding corrective context. Based on these results, we propose two lightweight mitigations that substantially reduce sycophancy while matching or exceeding memory systems at factual recall.