LM-Polygraph/gsm8k

这是一个为LM-Polygraph基准测试而预处理的gsm8k数据集版本。数据集包含continuation子集，对应于LM-Polygraph中的主数据集，可能还包含其他对应于指令方法的子集。每个子集都有训练集和测试集两个划分，每个划分包含两个字符串列：input和output，分别对应于LM-Polygraph的处理输入和处理输出。

This is a preprocessed version of the gsm8k dataset for benchmarks in LM-Polygraph. The dataset contains the continuation subset, which corresponds to the main dataset in LM-Polygraph, and may also contain other subsets corresponding to instruct methods. Each subset has two splits: train and test, with each split containing two string columns: input and output, which correspond to the processed input and output for LM-Polygraph, respectively.