alinia/safety_guard_v2

该数据集包含多语言文本数据，主要用于安全内容检测任务。数据集包含以下字段：id（唯一标识符）、prompt（提示文本）、response（响应文本）、label（标签）、is_unsafe（是否不安全，二分类）、language（语言）、source（数据来源）和is_unsafe_str（不安全文本描述）。数据集分为训练集（52,642条）和验证集（2,771条）。

This dataset contains multilingual text data, primarily used for safe content detection tasks. The dataset includes the following fields: id (unique identifier), prompt (prompt text), response (response text), label (label), is_unsafe (whether unsafe, binary classification), language (language), source (data source), and is_unsafe_str (unsafe text description). The dataset is divided into a training set (52,642 entries) and a validation set (2,771 entries).