CodeIsAbstract/reasoning_dataset

这是一个基于Dolphin R1数据集修改的文本分类数据集，专门用于识别文本中的推理内容。数据集整合了原始Dolphin R1数据集中的所有三个数据集，并提取了助手消息。每个数据点包括文本和标签两个字段，标签为二进制，表示文本是否包含推理。

This is a modified text classification dataset based on the Dolphin R1 dataset, specifically designed for identifying reasoning in text. The dataset combines all three datasets from the original Dolphin R1 release and extracts the assistant messages. Each data point consists of a text field and a label field, which is binary, indicating whether the text contains reasoning or not.