1 paper · 1 filter
Chiyu Zhang, Huiqin Yang, Bendong Jiang +8
The rapid proliferation of LLM-based autonomous agents in real operating system environments introduces a new category of safety risk beyond content safety: behavior jailbreak, whe…