Picture for Wenjun Xue

Wenjun Xue

Defending LLM-based Multi-Agent Systems Against Cooperative Attacks with Sentence-Level Rectification

Add code
May 27, 2026
Viaarxiv icon