LM-Polygraph/truthfulqa

这是一个为LM-Polygraph基准测试而创建的truthfulqa数据集的预处理版本。它包括两个子集：continuation和simple_instruct，每个子集都有训练集和测试集。每个split包含两个字符串字段：输入(input)和输出(output)。该数据集的目的是将数据集创建代码与基准测试代码分离。This is a preprocessed version of the truthfulqa dataset created for benchmarks in LM-Polygraph. It includes two subsets: continuation and simple_instruct, each with a training set and a test set. Each split contains two string fields: input and output. The purpose of this dataset is to separate the dataset creation code from the benchmarking code.