jplhughes2/alignment-faking-synthetic-chat-dataset-recall-0k-docs-8k-benign-2k-refusals
收藏数据链接：
官方服务：
资源简介：
该数据集包含两个字符串类型的特征：提示（prompt）和完成（completion）。它有一个训练集分割，包含大约10000个示例，数据集总大小约为12MB。数据集的具体内容和用途未在README中描述。
The dataset includes two string-type features: prompt and completion. It has a training split with approximately 10,000 examples, and the total size of the dataset is about 12MB. The specific content and purpose of the dataset are not described in the README.
提供机构：
jplhughes2
