windfromthenorth/crafter-world-sft-multi-split

该数据集包含四个字段：查询(query)、完成(completion)、预测范围(horizon)和令牌数量(num_tokens)。查询和完成字段为字符串类型，预测范围和令牌数量为整数类型。数据集分为训练集和测试集，其中训练集包含1,917,534个示例，测试集包含213,060个示例。数据集的下载大小为526,188,874字节，总大小为4,232,501,058字节。

The dataset includes four fields: query, completion, horizon, and num_tokens. The query and completion fields are of string type, while the horizon and num_tokens fields are of integer type. The dataset is divided into training and test sets, with the training set containing 1,917,534 examples and the test set containing 213,060 examples. The download size of the dataset is 526,188,874 bytes, and the total size is 4,232,501,058 bytes.