Code for "Semantically Equivalent Adversarial Rules for Debugging NLP Models" - View it on GitHub