MMODELYST
People/Bingyu Yan
PPL

Bingyu Yan

Papers (3)
Evo-Attacker: Memory-Augmented Reinforcement Learning for Long-Horizon Tool Attacks on LLM-MASModel-Agnostic Lifelong LLM Safety via Externalized Attack-Defense Co-EvolutionPropGuard: Safeguarding LLM-MAS via Propagation-Aware Exploration and Remediation