misinfo-general

We introduce misinfo-general, a benchmark dataset for evaluating misinformation models' ability to perform out-of-distribution generalisation. Misinformation changes rapidly, much quicker than moderators can annotate at scale, resulting in a shift between the training and inference data distributions. As a result, misinformation models need to be able to perform out-of-distribution generalisation, an understudied problem in existing datasets. In our paper, we identify 6 axes of generalisation-time, event, topic, publisher, political bias, misinformation type-and design evaluation procedures for each. We also analyse some baseline models, highlighting how these fail important desiderata.