Embedded Lies Dataset
收藏资源简介:
该数据集名为Embedded Lies Dataset,由IMT高级研究学院卢卡分校、蒂尔堡大学和伦敦大学学院的研究团队创建,旨在研究嵌入谎言在口头欺骗检测中的应用。数据集包含2088条真实和包含嵌入谎言的陈述,数据来源于1058名英语流利的参与者,通过在线平台Prolific招募。数据集的创建过程采用受试者内设计,参与者首先提供真实的自传体事件描述,然后重写这些陈述以包含嵌入的谎言,并标注谎言的中央性、欺骗性和来源。数据集的应用领域主要集中在自动化谎言检测,特别是针对嵌入谎言的检测,旨在解决真实与谎言混合的复杂欺骗场景中的检测难题。
This dataset, named Embedded Lies Dataset, was developed by a research team from IMT School for Advanced Studies Lucca, Tilburg University, and University College London, with the goal of investigating the application of embedded lies in verbal deception detection. The dataset consists of 2088 statements that are either truthful or contain embedded lies, sourced from 1058 fluent English-speaking participants recruited through the online platform Prolific. The dataset was constructed using a within-subjects experimental design: participants first provided truthful descriptions of autobiographical events, then rewrote these statements to incorporate embedded lies, and annotated the centrality, deceptiveness, and source of the embedded lies. The primary application scope of this dataset centers on automated deception detection, particularly the detection of embedded lies, aiming to address the detection challenges in complex deception scenarios where truthful and deceptive statements are mixed.




