1 paper
Jinru Ding, Chao Ding, Yidong Jiang +9
As large language models (LLMs) become high-privilege agents in risk-sensitive settings, they introduce systemic threats beyond hallucination, where minor compliance errors can cau…