@inproceedings{xiang-etal-2021-toxccin,
    title = "{T}ox{CCI}n: Toxic Content Classification with Interpretability",
    author = "Xiang, Tong  and
      MacAvaney, Sean  and
      Yang, Eugene  and
      Goharian, Nazli",
    booktitle = "Proceedings of the Eleventh Workshop on Computational Approaches to Subjectivity, Sentiment and Social Media Analysis",
    month = apr,
    year = "2021",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2021.wassa-1.1",
    pages = "1--12",
    abstract = "Despite the recent successes of transformer-based models in terms of effectiveness on a variety of tasks, their decisions often remain opaque to humans. Explanations are particularly important for tasks like offensive language or toxicity detection on social media because a manual appeal process is often in place to dispute automatically flagged content. In this work, we propose a technique to improve the interpretability of these models, based on a simple and powerful assumption: a post is at least as toxic as its most toxic span. We incorporate this assumption into transformer models by scoring a post based on the maximum toxicity of its spans and augmenting the training process to identify correct spans. We find this approach effective and can produce explanations that exceed the quality of those provided by Logistic Regression analysis (often regarded as a highly-interpretable model), according to a human study.",
}