AMT and Upwork Deepfake Text Detection Datasets

该数据集包含了来自非专业个体（亚马逊土耳其机器人参与者）和专业写作者（Upwork参与者）的回应，他们被委托识别深度伪造文本。在个人和协作环境下，参与者们在其准确检测深度伪造文本的能力上接受了评估。该数据集记录了参与者在个人和协作环境下的表现，为协作在识别深度伪造文本中的有效性提供了洞见。此外，该数据集包含了50个问题，用于准确性的评估，任务是对深度伪造文本的检测。

This dataset contains responses from non-professional individuals (Amazon Mechanical Turk workers) and professional writers (Upwork participants), who were tasked with identifying deepfake texts. Participants were evaluated on their ability to accurately detect deepfake texts in both individual and collaborative working environments. The dataset records participants' performance across these two settings, offering insights into the effectiveness of collaboration in deepfake text detection. Additionally, the dataset includes 50 questions developed for accuracy evaluation, with the core task being deepfake text detection.