1
Adversarial Feeds Steer LLM Agent Decisions Against Their Defaults
AI代理在消费信息流时易被上游排序器恶意操控，彻底颠覆默认决策，安全评估亟需覆盖排序环节
arXiv:2606.00914v1 Announce Type: new Abstract: LLM agents increasingly act after consuming ranked external information streams such as social feeds, …