rungao2001/tldr-llama32-processed-pref

该数据集包含四个主要特征：query（查询）、chosen（选择的答案）、rejected（拒绝的答案）和from（来源），所有特征的数据类型均为字符串。数据集分为训练集、验证集和测试集，训练集包含92,858个样本，验证集包含33,083个样本，测试集包含53,003个样本。数据集的下载大小为42,439,510字节，总大小为311,266,542字节。数据文件路径分别为：训练集（data/train-*）、验证集（data/valid-*）和测试集（data/test-*）。

The dataset contains four main features: query, chosen, rejected, and from, all of which are of string type. The dataset is divided into training, validation, and test sets, with the training set containing 92,858 samples, the validation set containing 33,083 samples, and the test set containing 53,003 samples. The download size of the dataset is 42,439,510 bytes, and the total size is 311,266,542 bytes. The data file paths are: training set (data/train-*), validation set (data/valid-*), and test set (data/test-*).