mhb-maaz/ai-detector-dataset

这是一个用于二分类任务的数据集，旨在区分AI生成的代码和人类编写的代码。数据集包含了500,000个训练样本，100,000个开发样本和10,000个测试样本，每个样本都是一个代码片段和对应的标签（人类编写或AI生成）。数据集支持TF-IDF、XGBoost和CodeBERT等模型的训练和评估。

This is a dataset designed for binary classification tasks, aiming to distinguish between AI-generated code and human-written code. The dataset contains 500,000 training samples, 100,000 development samples, and 10,000 test samples, each consisting of a code snippet and a corresponding label (human-written or AI-generated). The dataset supports the training and evaluation of models such as TF-IDF, XGBoost, and CodeBERT.