While pretrained language models ("LM") have driven impressive gains over morpho-syntactic and semantic tasks, their ability to model discourse and pragmatic phenomena is less clear. As a step towards a better understanding of their discourse modelling capabilities, we propose a sentence intrusion detection task. We examine the performance of a broad range of pretrained LMs on this detection task for English. Lacking a dataset for the task, we introduce INSteD, a novel intruder sentence detection dataset, containing 170,000+ documents constructed from English Wikipedia and CNN news articles. Our experiments show that pretrained LMs perform impressively in in-domain evaluation, but experience a substantial drop in the cross-domain setting, indicating limited generalisation capacity. Further results over a novel linguistic probe dataset show that there is substantial room for improvement, especially in the cross-domain setting.
@article{arxiv.2103.10133,
title = {Evaluating Document Coherence Modelling},
author = {Aili Shen and Meladel Mistica and Bahar Salehi and Hang Li and Timothy Baldwin and Jianzhong Qi},
journal= {arXiv preprint arXiv:2103.10133},
year = {2021}
}