mlfoundations-dev/stackoverflow_25000tasks_.5p

该数据集包含了指令、完成情况、对话信息（对话者及对话内容）、种子任务、生成的命令提示、生成的命令、命令响应、最终指令和最终响应等字段。数据集的训练集部分包含50000个样本，总文件大小为772366808字节。

The dataset includes fields such as instruction, completion, conversation information (including the interlocutor and the content of the conversation), seed tasks, generated command prompts, generated commands, command responses, final instructions, and final responses. The training set of the dataset contains 50,000 samples with a total file size of 772366808 bytes.