arXiv · Software Engineering· Jeremy Spence, Nicholas Assaderaghi, Feng Xiao, Jinhao Zhu, Nikil Ravi, Xiangyu Qi, Matthew Jagielski, Raluca Ada Popa, Eric Wallace, Guannan Wei, Yangruibo Ding, Zhuo Zhang·· 3 hr agoSelectedAI score60
SRE-Bench:首个无污染的真实逆向工程基准
The Next Challenge for Agentic Cybersecurity: A Realistic, Contamination-Free Reverse Engineering Benchmark
AI brief
论文提出 SRE-Bench,由逆向工程专家耗时逾 5,000 小时构建,包含 19 个真实规模程序、262 个二进制实例和 1,572 个确定性评分任务。
Why it matters
真实二进制逆向工程对前沿 Agent 仍具挑战,来源代码安全能力未必迁移到二进制分析。
Source: arXiv · Software Engineering · arxiv.org