mlfoundations-dev/OH_DCFT_v3_wo_airoboros

该数据集包含对话数据，每个对话包含来源和内容值两个主要部分，以及一个精确的源标签序列。数据集分为训练集，包含966,003个示例，总大小为1,919,987,704字节。

This dataset contains conversation data, each consisting of a source and a content value, along with an exact source label sequence. The dataset is divided into a training set, which includes 966,003 examples with a total size of 1,919,987,704 bytes.