Numeration

该数据集旨在为数值任务中的每个数字表示独特的词汇，同时关注插值和 extrapolation（外推）设置。在模型训练中，该数据集采用了分割数字和完整整数表示法。规模上，训练集达到了4.9千条数据，任务专注于数字表示。

This dataset is designed to assign unique lexical representations to each numeral in numerical tasks, while focusing on both interpolation and extrapolation settings. For model training, this dataset adopts two representation formats: digit-separated and full integer notations. In terms of scale, the training set contains 4.9 thousand data samples, with the task focusing on numeral representation.