Large language model (LLM) agents are increasingly used for vulnerability analysis, yet existing benchmarks provide limited coverage of exploitation that dep…
机构:清华
来源:arXiv 2610.17837 | AI4Papers 论文推荐平台