Holmeister/OffensEval-multi

该数据集包含五个主要字段：instruction（指令）、input（输入）、output（输出）、inst_no（指令编号）和system（系统）。数据集分为训练集、验证集和测试集，分别包含10000、500和5000个样本。训练集大小为4229512字节，验证集大小为208938字节，测试集大小为2053910字节。总数据集大小为6492360字节，下载大小为1352847字节。数据文件路径分别为：训练集（data/train-*）、验证集（data/validation-*）和测试集（data/test-*）。

The dataset contains five main fields: instruction, input, output, inst_no, and system. The dataset is divided into training, validation, and test sets, containing 10,000, 500, and 5,000 samples respectively. The training set size is 4,229,512 bytes, the validation set size is 208,938 bytes, and the test set size is 2,053,910 bytes. The total dataset size is 6,492,360 bytes, and the download size is 1,352,847 bytes. The data file paths are: training set (data/train-*), validation set (data/validation-*), and test set (data/test-*).