DAMO-NLP-SG/Mistral-7B-LongPO-128K-tokenized
收藏Hugging Face2025-02-17 更新2025-04-12 收录
下载链接：
https://hf-mirror.com/datasets/DAMO-NLP-SG/Mistral-7B-LongPO-128K-tokenized
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含文本样本的特征，其中包括选中的输入ID、注意力掩码和标签，以及被拒绝的样本的相应特征。同时还包括这些样本的参考特征。数据集仅包含训练集，共有约44850个样本，数据大小约为359.44GB。
The dataset consists of text sample features, including chosen input IDs, attention masks, and labels, as well as the corresponding features for the rejected samples. It also includes reference features for these samples. The dataset only contains a training set with approximately 44850 samples, totaling about 359.44GB in size.
提供机构：
DAMO-NLP-SG
