Leakage-Adjusted Simulatability: Can Models Generate Non-Trivial Explanations of Their Behavior in Natural Language?

Authors:
  • PHPeter Hase
    Peter Hase
  • SZShiyue Zhang
    Shiyue Zhang
  • HXHarry Xie
    Harry Xie
  • MBMohit Bansal
    Mohit Bansal

Abstract
Peter Hase, Shiyue Zhang, Harry Xie, Mohit Bansal. Findings of the Association for Computational Linguistics: EMNLP 2020. 2020.