withmartian/binary_toxic

这是一个包含文本数据的训练集，其中包括提示(prompt)、元信息(meta)、响应(response)等字段。元信息中包含了生成信息(generation)、生成方法(generation_method)、分组(group)等子字段。此外，还包括属性名称(attribute_name)和属性值(attribute_value)，以及数据来源(source)。训练集部分包含超过25万条示例，文件大小约为340MB。

This is a training dataset containing text data, which includes fields such as prompt, meta, and response. The meta information contains sub-fields like generation, generation method, and group. In addition, it includes attribute name (attribute_name) and attribute value (attribute_value), as well as the data source (source). The training set contains more than 250,000 examples, with a file size of approximately 340MB.