This repository contains code and data for the paper "Concept-Based Explanations to Test for False Global Causal Relationships Learned by Abusive Language Classifiers" Published at WOAH2023 Co-located with ACL2023. The paper can be found at https://www.svkir.com/papers/Nejadgholi-et-al-TCAVcausal-WOAH-2023.pdf.

For more information, please refer to the paper.

Contact:

isar.Nejadgholi@nrc-cnrc.gc.ca

Terms of Use:

The code can be used freely for non-commercial research and educational purposes.

Do not redistribute the code.

National Research Council Canada (NRC) disclaims any responsibility for the use of the codes listed here and does not provide technical support.