neko-llm/HLE_DPO_HarmfulQA
收藏Hugging Face2025-08-02 更新2025-08-09 收录
下载链接:
https://hf-mirror.com/datasets/neko-llm/HLE_DPO_HarmfulQA
下载链接
链接失效反馈官方服务:
资源简介:
该数据集包含一个问题及其两个可能的输出(偏好输出和非偏好输出)。它适用于训练模型来识别和生成特定问题的首选回答。
This dataset contains a question and two possible outputs for it (preferred output and non-preferred output). It is suitable for training models to identify and generate preferred responses to specific questions.
提供机构:
neko-llm



