VulDetectBench

该数据集旨在通过五个难度逐渐增加的任务，评估大型语言模型（LLMs）在漏洞检测方面的能力。它不仅对LLMs在识别、分类和定位漏洞方面的能力进行了评估，而且还提供了LLMs在不同任务中的性能洞察。该数据集的任务重点在于漏洞的检测与分析。

This dataset is designed to evaluate the vulnerability detection capabilities of Large Language Models (LLMs) through five tasks with gradually increasing difficulty. It not only assesses LLMs' abilities in vulnerability identification, classification and localization, but also provides performance insights of LLMs across different tasks. The tasks of this dataset focus on vulnerability detection and analysis.