有人做了个叫 MemTrapBench 的基准，专门测大模型记忆里的认知陷阱。他们发现，记忆反而会拖累模型表现，连最好的方法都掉10%以上。他们还搞了个叫 AdaptiveMem 的方法来解决这个问题。
研究人员提出 MemTrapBench 基准，用于评估大模型记忆使用中的认知陷阱。该基准涵盖 Reasoning Fixation 和 Belief Distortion 两种陷阱形式。实验显示，在两个模型家族和五种代表性记忆框架上，所有记忆策略的表现均不如无记忆设置，最强方法性能下降超10%。为缓解这些陷阱，研究者提出 AdaptiveMem 方法，可在推理时指导模型避免记忆陷阱。
MemTrapBench: Benchmarking Cognitive Traps in LLM Memory Use
Memory has become a key component of large language models, enabling them to retain information and learn from long-term interactions. However, existing memory benchmarks mainly evaluate whether information is correctly extracted, stored, and retrieved, while largely overlooking how retrieved memories reshape model reasoning and affect performance on the current task. We identify memory-induced cognitive traps: even faithfully recorded and semantically relevant memories can distort model reasoning or beliefs and degrade current task performance. To systematically evaluate these failure modes, we introduce MemTrapBench, which covers two forms of cognitive traps: Reasoning Fixation and Belief Distortion. Experiments across two model families and five representative memory frameworks show that MemTrapBench is challenging: all evaluated memory strategies underperform the no-memory setting, with even the strongest methods suffering drops of more than 10%. To mitigate these cognitive traps, we propose AdaptiveMem, a simple yet effective inference-time method that instructs LLMs to avoid memory traps. AdaptiveMem mitigates cognitive traps on MemTrapBench while preserving or improving performance on standard memory benchmarks across diverse memory frameworks.