arXivZhishang Xiang, Zerui Chen, Yunbo Tang, Zhimin Wei, Ruqin Ning, Yujie Lin, Qinggang Zhang, Jinsong SuWed, Jul 1, 2026, 8:30 AM PDT
score 17.1
New benchmark tests AI agents for blind obedience to user memories
Original: MemSyco-Bench: Benchmarking Sycophancy in Agent Memory
Source: arxiv.org ↗
Writing ELI5 summary…