timaeus/pile-cc_loss_delta_ablation_l0h3
收藏Hugging Face2025-03-18 更新2025-04-12 收录
下载链接：
https://hf-mirror.com/datasets/timaeus/pile-cc_loss_delta_ablation_l0h3
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含文本和元数据两个特征，文本为字符串类型，元数据中包含pile_set_name字段，也是字符串类型。数据集有训练集 splits，包含10000个示例，总大小为87132301字节。数据集提供了一个默认配置，用于指定训练数据文件的路径。
The dataset includes two main features: text and metadata. The text feature is of string type, and the metadata feature contains a sub-feature pile_set_name, which is also of string type. The dataset is split into a training set (train) containing 10,000 examples with a total size of 87132301 bytes. The dataset provides a default configuration for specifying the path to the training data files.
提供机构：
timaeus
