Yusser/m_low_res_sae_wiki_big_tokenized_Llama-3.1-8B
收藏Hugging Face2025-08-05 更新2025-10-25 收录
下载链接：
https://hf-mirror.com/datasets/Yusser/m_low_res_sae_wiki_big_tokenized_Llama-3.1-8B
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含名为input_ids的整数序列特征，总共分为训练集(train)一个部分，训练集包含大约2284万个示例，数据集总大小约为93.6TB。数据集的下载大小为47.2GB。
The dataset includes a feature named input_ids consisting of integer sequence values. It is split into one part, the training set (train), containing approximately 22.84 million examples, with the total dataset size being approximately 93.6TB. The download size of the dataset is 47.2GB.
提供机构：
Yusser
