FineFake is a dataset specifically designed for fine-grained, multi-domain fake news detection, jointly created by Beijing University of Aeronautics and Astronautics and Beijing University of Posts and Telecommunications. The dataset contains 16,909 data samples, covering 6 semantic topics and 8 different platforms. Each news sample contains various forms of content, including text, images, and potential social context information, and undergoes semi-manual verification to validate common knowledge.

Unlike traditional binary labels for true and false news, the FineFake dataset provides more granular classification, helping to more accurately reveal the strategies behind fake news. FineFake aims to address the domain adaptability problem in fake news detection by providing cross-topic and platform data, enabling researchers to develop detection models that can accurately identify and adapt to different news domains.