SICK

该数据集包含了用于评估语义相似度的句子对及其相似度得分，它被用于评价语言模型在学习上下文语言表示方面的能力。该数据集的规模由一系列的句子对组成，旨在完成无监督的语义文本相似度任务。

This dataset contains sentence pairs and their corresponding similarity scores for evaluating semantic similarity, and it is employed to assess the capability of language models in learning contextual language representations. Composed of a series of sentence pairs, this dataset is designed for unsupervised semantic textual similarity tasks.