To what extent do human explanations of model behavior align with actual model behavior?

Grusha Prasad, Yixin Nie, Mohit Bansal, Robin Jia, Douwe Kiela, Adina Williams

BibTeX
@inproceedings{prasad2021what,
  author = {Grusha Prasad and Yixin Nie and Mohit Bansal and Robin Jia and Douwe Kiela and Adina Williams},
  title = {{To what extent do human explanations of model behavior align with actual model behavior?}},
  booktitle = {BlackboxNLP @ EMNLP},
  year = {2021},
  doi = {10.18653/v1/2021.blackboxnlp-1.1},
}
Figure