The ability to generalise is one of the primary desiderata of natural language processing (NLP). Yet, how "good generalisation" should be defined and evaluated is not well understood, nor are there common standards to evaluate generalisation. As a consequence, newly proposed models are usually not systematically tested for their ability to generalise. GenBench's mission is to make state-of-the-art generalisation testing the new status-quo in NLP. As a first step, we present a generalisation taxonomy, describing the underlying building blocks of generalisation in NLP. We use the taxonomy to do an elaborate review of over 400 generalisation papers, and we make recommendations for promising areas for the future. You can find all this information in our taxonomy paper. On this website, you can learn about the taxonomy, you can visually explore our results, and get citations from our analysis or contribute papers that we will periodically add to our review.


December 6, 2022 GenBench presentation by Karim Lasri @ ENS
November 14, 2022 GenBench presentation by Christos Christodoulopoulos @Amazon
November 10, 2022 We are working on a *ACL workshop proposal, express your interest in our twitter poll, stay tuned for more info.
November 2, 2022 GenBench presentation by Dieuwke Hupkes @Deep Learning & AI Talks hosted by Qualcomm Amsterdam
October 7, 2022 We are live! Read a summary tweet about our taxonomy paper, browse our website to explore the results, and stay tuned!