ConFiguRe

该数据集是一个面向情境化图形识别的中文语料库，旨在从话语层面的语境中提取图形单元，并将它们归类到适当的图形类型中。该数据集包含中文中常用的12种图形类型，并作为测试当前最先进的比喻语言理解技术的基准。尽管规模较小且分布不均，该数据集仍包含1757个预测，任务涵盖了图形提取、图形类型分类和图形识别。

This dataset is a Chinese corpus for contextualized figure-of-speech recognition, designed to extract figure units from discourse-level context and classify them into appropriate figure-of-speech categories. It covers 12 commonly used figure-of-speech categories in Chinese, and serves as a benchmark for evaluating state-of-the-art figurative language understanding technologies. Despite its small scale and uneven distribution, the dataset contains 1757 instances, with tasks including figure extraction, figure category classification, and figure recognition.