paolordls/crosslg-news-sm

该数据集包含文本内容、真实性标签、文章ID和片段ID等特征。数据集分为unlearn、train、forget和retain四个部分，每个部分都有对应的字节大小和样本数量。数据文件路径分别为data/train-*、data/unlearn-*、data/retain-*和data/forget-*。

The dataset includes features such as text content, real/fake labels, article IDs, and snippet IDs. The dataset is divided into four parts: unlearn, train, forget, and retain, each with corresponding byte sizes and sample counts. The data file paths are data/train-*, data/unlearn-*, data/retain-*, and data/forget-*.