BEE-spoke-data/LONGCOT-merged-1M

该数据集是由PowerInfer/QWQ-LONGCOT-500K和PowerInfer/LONGCOT-Refine-500K两个数据集混合而成的文本生成数据集。数据集经过处理，移除了无意义的起始短语，并且对en配置进行了筛选，只包含检测为英语的对话行。数据集包含prompt和response两个字段，都是字符串类型，分为训练集，提供了详细的大小和示例数量信息。

This dataset is a text generation dataset created by combining PowerInfer/QWQ-LONGCOT-500K and PowerInfer/LONGCOT-Refine-500K. The dataset has been processed to remove meaningless starting phrases and the en configuration has been filtered to include only dialogue lines detected as English in both the prompt and response columns. The dataset contains two fields, prompt and response, both of which are strings, and is divided into a training set with detailed information on size and the number of examples provided.