"Dolores An LLM-Driven Debugging Agent for Root-Cause Analysis"
收藏DataCite Commons2026-04-02 更新2026-05-03 收录
下载链接：
https://ieee-dataport.org/documents/dolores-llm-driven-debugging-agent-root-cause-analysis
下载链接
链接失效反馈官方服务：
资源简介：
"This is a dataset for vulnerability cause analysis, which can be tested using debugging methods. This dataset consists of 50 vulnerabilities from 29 software systems, including 42 officially assigned CVEs. The selected vulnerabilities cover six common types: integer overflow, heap buffer overflow, stack buffer overflow, use-after-free, double-free, and null pointer dereference. In order to preserve the validity of CVEs related ground truth (i.e., avoiding incorrect attributions of rootcauses), we primarily selected vulnerabilities disclosed around2023. The cases vary significantly in size, ranging from 796 to 2,270,065 lines of code, with various complexity levels. As the benchmarks are predominantly overflow-related, we partially use Address Sanitizer (ASAN) to instrument programs to catch crashes."
提供机构：
IEEE DataPort
创建时间：
2026-04-02
