ChamaraVishwajithRajapaksha/Code-Vulnerability-FineTune

该数据集是一个专门用于微调大型语言模型（LLMs）进行软件漏洞检测和分析的对话数据集。它基于`ChamaraVishwajithRajapaksha/Code_Vulnerability_Dataset`预处理和转换而来，采用ShareGPT/FineTome对话格式。数据集包含C/C++代码样本，每个样本包含两轮对话：人类提问（代码分析请求）和AI回答（漏洞分析）。数据集覆盖了多种CWE漏洞类型，如CWE-787（越界写入）、CWE-416（释放后使用）等。适用于安全代码审查、漏洞检测模型训练、代码感知安全助手构建等用途。

This dataset is a conversation dataset specifically designed for fine-tuning large language models (LLMs) for software vulnerability detection and analysis. It is preprocessed and transformed from the `ChamaraVishwajithRajapaksha/Code_Vulnerability_Dataset` and adopts the ShareGPT/FineTome conversation format. The dataset includes C/C++ code samples, each containing two turns of conversation: a human question (code analysis request) and an AI answer (vulnerability analysis). The dataset covers various CWE vulnerability types, such as CWE-787 (Out-of-bounds Write), CWE-416 (Use After Free), etc. It is suitable for security code review, vulnerability detection model training, building code-aware security assistants, and other applications.