QuanTemp

QuanTemp是由代尔夫特理工大学等机构创建的一个大型、多领域的真实世界数据集，专注于验证数值声明。该数据集包含15,514个涉及比较、统计、区间和时间方面的数值声明，以及详细的元数据和伴随的证据收集。数据集的创建过程涉及从多个事实检查组织收集真实声明，并通过严格的筛选和证据收集确保数据质量。QuanTemp主要用于自动化事实检查系统的发展，特别是在处理复杂和信息不精确的数值声明方面，旨在提高事实检查的准确性和效率。

QuanTemp is a large-scale, multi-domain real-world dataset dedicated to numerical claim verification, developed by Delft University of Technology and other institutions. This dataset includes 15,514 numerical claims covering comparison, statistical, interval, and temporal dimensions, along with detailed metadata and accompanying evidence collections. The dataset construction process involved collecting authentic claims from multiple fact-checking organizations, and ensuring data quality through rigorous filtering and evidence gathering protocols. QuanTemp is primarily intended for the development of automated fact-checking systems, particularly for addressing complex and informationally imprecise numerical claims, with the objective of improving the accuracy and efficiency of fact-checking practices.