mteb/CDSC-R

CDSC-R是一个用于文本相关性评估的语料库，包含两个句子及其相似度评分，用于评估分布语义模型的组合性。数据集覆盖网络和书面领域，分为训练集、验证集和测试集，适用于句子相似度评分和自然语言推理任务。

CDSC-R is a corpus for textual relatedness evaluation, containing two sentences and their similarity scores, used for assessing the compositionality of distributional semantics models. The dataset covers web and written domains and is split into training, validation, and test sets, suitable for sentence similarity scoring and natural language inference tasks.