Kyleyee/train_data_HH_explicit_prompt
收藏Hugging Face2025-03-14 更新2025-04-12 收录
下载链接：
https://hf-mirror.com/datasets/Kyleyee/train_data_HH_explicit_prompt
下载链接
链接失效反馈官方服务：
资源简介：
HH-RLHF-Helpful-Base数据集是一个专门为偏好学习和对齐任务而设计的处理过的数据集，它包含根据人类评估者的偏好标记为有用（chosen）或不太有用（rejected）的文本样本对，用于训练模型生成对用户更有帮助的回应。
The HH-RLHF-Helpful-Base dataset is a curated version of the Anthropic HH-RLHF dataset, designed for preference learning and alignment tasks. It includes pairs of text samples labeled as chosen or rejected based on human evaluators preferences for helpfulness, aimed at training models to generate more helpful responses for users.
提供机构：
Kyleyee
