seangogo/processed_tldr_comparison_dataset_20251102_065554
收藏数据链接:
官方服务:
资源简介:
这是一个用于训练奖励模型的比较数据集。每个样例包含一个用于摘要的查询(帖子)和两个响应(选择的和被拒绝的),其中选择的响应是通过人类反馈更受欢迎的。
This is a comparison dataset used for training a reward model. Each example contains a query (post) and two responses (chosen and rejected) where the chosen response is preferred by human feedback.
提供机构:
seangogo


