遇见数据集

Metaskepsis/orpo_medium

收藏
Hugging Face2025-01-19 更新2025-04-12 收录
官方服务:

资源简介:

这是一个包含多个字段的数据集,用于某种类型的问题回答任务。数据集的字段包括问题类型、问题内容、正确答案、提示信息(包括内容和角色)、选中答案(包括内容和角色)、被拒绝答案(包括内容和角色)、选中答案得分、被拒绝答案得分、数据ID和对齐信息。数据集被划分为训练集,共有86464个示例,大小为426124138字节。

This is a dataset with multiple fields designed for a question answering task of some type. The dataset fields include the type of problem, the content of the problem, the correct answer, prompt information (including content and role), chosen answer (including content and role), rejected answer (including content and role), score of the chosen answer, score of the rejected answer, data ID, and alignment information. The dataset is split into a training set, which contains 86464 examples and is 426124138 bytes in size.

提供机构:
Metaskepsis
二维码
社区交流群
二维码
科研交流群
商业服务