同伴记忆改变 LLM 智能体咨询对象，但未检测到答案准确率提升
原标题：Peer Memory Changes Who LLM Agents Consult Without a Detected Gain in Answer Accuracy
AI 摘要
该研究提出 PeerMem，将 MACE 中用于同伴选择的 LinUCB 状态持久化保存，并在相同智能体再次出现时复用。实验使用六个智能体、80 道 HotpotQA 问题，对比冷启动、热启动与打乱初始化三种条件，发现热记忆降低了事后遗憾并增加了对其他证据持有者的咨询，但最终答案 F1 的差异仍不确定。在 Qwen3.5-9B 上进一步测试角色重分配场景，主要交互效应为 -1.38 分，95% 区间 [-3.50, +0.73]。
正文节选
Personalized, Aligned, Long-Term Memory for AI Systems (PALM) Peer Memory Changes Who LLM Agents Consult Without a Detected Gain in Answer Accuracy Abstract Multi-agent LLM systems often reuse the same agents across tasks, creating an opportunity to draw on past interactions when deciding whom to consult. We study this form of memory with PeerMem, which saves MACE’s LinUCB state and reuses it when the same agents return. Using six agents, a fixed set of 80 HotpotQA questions, and 12 paired linea