SecurityEval

该数据集包含了121个Python样本，每一个都与69种常见弱点枚举（CWE）类型中的某一漏洞相关联。这样的结构使得它成为评估在大型语言模型生成的代码中安全性的理想选择。此外，该数据集还用于评估生成代码的安全性，并有助于识别潜在的安全漏洞。该数据集的规模为121个样本，其任务是对生成代码的安全性进行评估。

This dataset contains 121 Python code samples, each associated with a vulnerability belonging to one of the 69 Common Weakness Enumeration (CWE) categories. This structure makes it an ideal candidate for evaluating the security of code generated by Large Language Models (LLMs). Additionally, this dataset is utilized for assessing the security of generated code and aids in identifying potential security vulnerabilities. With a total of 121 samples, this dataset is tailored for the task of evaluating the security of generated code.

搜集汇总

背景与挑战

背景概述

SecurityEval是一个包含121个Python样本的数据集，每个样本都与69种常见弱点枚举（CWE）类型中的特定漏洞相关联，专门用于评估大型语言模型生成代码的安全性，并帮助识别潜在安全漏洞。