timaeus/min_loss_delta_ablation_l0h1

该数据集包含三个主要特征：contents（内容，数据类型为字符串）、metadata（元数据，包含一个名为pile_set_name的字符串序列）和id（标识符，数据类型为int64）。数据集有一个名为train的分割，包含10,000个样本，总大小为17,536,478字节。下载大小为7,419,565字节，数据集总大小为17,536,478字节。配置部分指定了默认配置，数据文件路径为data/train-*。

The dataset contains three main features: contents (content, data type is string), metadata (metadata, containing a string sequence named pile_set_name), and id (identifier, data type is int64). The dataset has a split named train containing 10,000 samples, with a total size of 17,536,478 bytes. The download size is 7,419,565 bytes, and the total dataset size is 17,536,478 bytes. The configuration section specifies the default configuration, with the data file path being data/train-*.