marcov/hyperpartisan_news_detection_bypublisher_promptsource

该数据集包含多个特征，如文本、标题、超党派性、URL、发布日期、偏见、模板名称、模板、渲染输入和渲染输出。偏见特征是一个分类标签，包含从极右到极左的不同政治倾向。数据集分为训练集和验证集，分别包含300万和75万条记录。数据集的下载大小和总大小也提供了具体数值。

This dataset includes multiple features such as text, title, hyperpartisan, URL, published date, bias, template name, template, rendered input, and rendered output. The bias feature is a class label that includes different political orientations from far-right to far-left. The dataset is divided into training and validation sets, containing 3 million and 750 thousand records respectively. The download size and total size of the dataset are also provided with specific values.