2020
Reducing Sentiment Bias in Language Models via Counterfactual Evaluation.
Proceedings of the Findings of the Association for Computational Linguistics: EMNLP 2020, 2020