Recalling Too Well: Sycophancy Evaluation and Mitigation in Memory-Augmented Models 事件

Recalling Too Well: Sycophancy Evaluation and Mitigation in Memory-Augmented Models arXiv:2606.10949v1 Announce Type: new Abstract: 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