FixMyPose

该数据集包含了从知名Wikidata人物中收集的负面陈述候选语句，旨在探讨这些语句在实体摘要中的实用性。此外，数据集还包括了来自三位众包工作者对陈述语句是否应被包含在传记摘要中的相关性注释。规模上，该数据集包含了2000条陈述（每个实体100条，每条实体20个候选）。任务是对实体摘要中的负面陈述进行排名。

This dataset comprises candidate negative statements collected from prominent Wikidata person entities, with the objective of investigating the utility of such statements in entity summaries. Additionally, the dataset includes relevance annotations from three crowdsourced workers regarding whether the statements should be included in biographical summaries. In terms of scale, the dataset contains 2000 statements: 100 statements per entity, with 20 candidates per statement. The task is to rank negative statements within entity summaries.