多语言软件漏洞检测数据集
收藏资源简介:
本研究构建了一个包含Python、Java和JavaScript三种流行编程语言漏洞数据的数据集。数据集由新加坡管理大学提供,包含了8260个Python语言的漏洞函数、7505个Java语言的漏洞函数和28983个JavaScript语言的漏洞函数。这些数据来源于国家漏洞数据库,并通过Git库中的漏洞修复提交记录进行提取和标注。该数据集旨在为大型语言模型在多语言软件漏洞检测任务上的性能评估提供基准,以解决当前软件安全领域中对不同编程语言漏洞检测方法效果缺乏全面研究的现状。
This study constructs a dataset containing vulnerability data for three popular programming languages: Python, Java, and JavaScript. Provided by the Singapore Management University, the dataset includes 8260 vulnerable functions written in Python, 7505 in Java, and 28983 in JavaScript. These data are sourced from the National Vulnerability Database and were extracted and labeled using vulnerability fix commit records in Git repositories. This dataset is intended to serve as a benchmark for evaluating the performance of Large Language Models (LLMs) on multilingual software vulnerability detection tasks, aiming to address the current lack of comprehensive research on the effectiveness of vulnerability detection methods across different programming languages in the software security domain.




