marcuscedricridia/unAIthical-ShareGPT
收藏Hugging Face2025-04-03 更新2025-04-12 收录
下载链接：
https://hf-mirror.com/datasets/marcuscedricridia/unAIthical-ShareGPT
下载链接
链接失效反馈官方服务：
资源简介：
unAIthical-ShareGPT数据集是一个合并了多个数据源的数据集，包括无警告过滤极性、有毒内容全过滤极性、毒性DPO过滤极性、有害内容过滤极性和不道德问答过滤极性数据集。该数据集遵循ShareGPT格式，每个会话包含对话的轮流发言，适用于进一步的自然语言处理研究和应用。
The unAIthical-ShareGPT dataset is a merged dataset from multiple sources, including no-warning filtered polarity, toxic content full filtered polarity, toxi-DPO filtered polarity, harmful content filtered polarity, and amoral QA filtered polarity datasets. The dataset follows the ShareGPT format, with each conversation containing turns of dialogue, suitable for further NLP research and applications.
提供机构：
marcuscedricridia
