HFXM/ShareGPT_safety_prompts

该数据集包含10,000个训练样本，每个样本包含一个prompt字段和一个relevance_score字段。prompt字段的数据类型为字符串，relevance_score字段的数据类型为浮点数。数据集的总大小为3,926,056字节，下载大小为2,311,437字节。数据集被配置为默认配置，数据文件路径为data/train-*。

The dataset contains 10,000 training samples, each with a prompt field and a relevance_score field. The prompt field is of string type, and the relevance_score field is of float64 type. The total size of the dataset is 3,926,056 bytes, with a download size of 2,311,437 bytes. The dataset is configured with a default configuration, and the data file path is data/train-*.