Tiamz/uplimit-synthetic-data-week-2-with-evol
收藏Hugging Face2025-04-05 更新2025-04-12 收录
下载链接:
https://hf-mirror.com/datasets/Tiamz/uplimit-synthetic-data-week-2-with-evol
下载链接
链接失效反馈官方服务:
资源简介:
该数据集是一个合成的数据集,用于指令微调,包含了生成的指令、被拒绝的响应、元数据(包括输入和输出令牌的数量)、模型名称和选择的文本。数据集通过distilabel工具生成,并可以重现生成管道。数据集目前只有训练集部分。
This dataset is a synthetic dataset for instruction tuning, containing generated instructions, rejected responses, metadata (including the number of input and output tokens), model name, and chosen text. The dataset is generated using the distilabel tool, and the generation pipeline can be reproduced. Currently, the dataset only includes the training set.
提供机构:
Tiamz



