CVEFixes
收藏资源简介:
CVEFixes数据集是由隆德大学的研究团队创建的,包含了从2021年6月9日到2024年7月23日的所有已发布CVE记录,涵盖了5365个CVE记录和1754个开源项目。数据集包含了5495个漏洞修复提交,并与相应的CVE记录关联,提供了CVE-ID、参考链接、严重性评分、CWE类型等信息。数据集主要用于研究语言模型在不同编程语言中的漏洞检测效果,特别是JavaScript、PHP、Java、Python和Go。通过对该数据集的分析,研究团队旨在评估语言模型在不同编程语言中的漏洞检测性能,并探讨代码复杂度与检测性能之间的关系。
The CVEFixes dataset was created by a research team from Lund University. It contains all publicly disclosed CVE records from June 9, 2021 to July 23, 2024, covering 5,365 CVE records across 1,754 open-source projects. The dataset includes 5,495 vulnerability fix commits associated with their corresponding CVE records, and provides information such as CVE-ID, reference links, severity scores, and CWE types. It is primarily used to study the vulnerability detection performance of language models across different programming languages, particularly JavaScript, PHP, Java, Python, and Go. Through analyses of this dataset, the research team aims to evaluate the vulnerability detection performance of language models across various programming languages and investigate the relationship between code complexity and detection performance.

- 1Vulnerability Detection in Popular Programming Languages with Language Models隆德大学 · 2024年



