mlsanityData to 14 Oct 2026

Contextualizing Hate Speech Classifiers with Post-hoc Explanation

arXiv 2005.02439 ↗First posted 5 May 2020cs.CL cs.IR cs.LGcited by 84

Not on the map. Why?

Abstract (arXiv)

Hate speech classifiers trained on imbalanced datasets struggle to determine if group identifiers like "gay" or "black" are used in offensive or prejudiced ways. Such biases manifest in false positives when these identifiers are present, due to models' inability to learn the contexts which constitute a hateful usage of identifiers. We extract SOC post-hoc explanations from fine-tuned BERT classifiers to efficiently detect bias towards identity terms. Then, we propose a novel regularization technique based on these explanations that encourages models to learn from the context of group identifiers in addition to the identifiers themselves. Our approach improved over baselines in limiting false positives on out-of-domain data while maintaining or improving in-domain performance. Project page: https://inklab.usc.edu/contextualize-hate-speech/.

Key facts. Cited by 84 papers in the map (data through 14 Oct 2026). First posted on arXiv 5 May 2020.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2021: 20; 2022: 23; 2023: 18; 2024: 11; 2025: 9; 2026: 3.

84 papers.

Sort:

By area:

  1. 2023XSTest: A Test Suite for Identifying Exaggerated Safety Behaviours in Large Language Modelscited by 505
  2. 2021A Survey on Gender Bias in Natural Language Processingcited by 89
  3. 2021On Measures of Biases and Harms in NLPcited by 81
  4. 2021A Survey of Race, Racism, and Anti-Racism in NLPcited by 75
  5. 2021Towards generalisable hate speech detection: a review on obstacles and solutionscited by 66
  6. 2022Handling Bias in Toxic Speech Detection: A Surveycited by 61
  7. 2022Entropy-based Attention Regularization Frees Unintended Bias Mitigation from Listscited by 55
  8. 2023Nationality Bias in Text Generationcited by 50
  9. 2023SemEval-2023 Task 10: Explainable Detection of Online Sexismcited by 48
  10. 2023HateMM: A Multi-Modal Dataset for Hate Video Classificationcited by 42
  11. 2022On Explaining Multimodal Hateful Meme Detection Modelscited by 29
  12. 2023Sociodemographic Bias in Language Models: A Survey and Forward Pathcited by 28
38 more
  1. 2021Refining Language Models with Compositional Explanationscited by 27
  2. 2024Recent Advances in Hate Speech Moderation: Multimodality and the Role of Large Modelscited by 26
  3. 2022An Empirical Study on Explanations in Out-of-Domain Settingscited by 21
  4. 2022Challenges in Applying Explainability Methods to Improve the Fairness of NLP Modelscited by 20
  5. 2021Fairness-aware Class Imbalanced Learningcited by 19
  6. 2021Mitigating Racial Biases in Toxic Language Detection with an Equity-Based Ensemble Frameworkcited by 19
  7. 2021Detecting Community Sensitive Norm Violations in Online Conversationscited by 18
  8. 2021Enjoy the Salience: Towards Better Transformer-based Faithful Explanations with Word Saliencecited by 16
  9. 2023Unmasking Nationality Bias: A Study of Human Perception of Nationalities in AI-Generated Articlescited by 16
  10. 2022KOLD: Korean Offensive Language Datasetcited by 14
  11. 2021Fight Fire with Fire: Fine-tuning Hate Detectors using Large Samples of Generated Hate Speechcited by 13
  12. 2021Identification of Bias Against People with Disabilities in Sentiment Analysis and Toxicity Detection Modelscited by 13
  13. 2021Hate Speech Classifiers Learn Human-Like Social Stereotypescited by 12
  14. 2022VisFIS: Visual Feature Importance Supervision with Right-for-the-Right-Reason Objectivescited by 12
  15. 2022Counterfactually Augmented Data and Unintended Bias: The Case of Sexism and Hate Speech Detectioncited by 11
  16. 2022Explainable Abuse Detection as Intent Classification and Slot Fillingcited by 11
  17. 2023Should We Attend More or Less? Modulating Attention for Fairnesscited by 11
  18. 2021Improving Counterfactual Generation for Fair Hate Speech Detectioncited by 10
  19. 2022Multilingual Auxiliary Tasks Training: Bridging the Gap between Languages for Zero-Shot Transfer of Hate Speech Detection Modelscited by 10
  20. 2024ToxVidLM: A Multimodal Framework for Toxicity Detection in Code-Mixed Videoscited by 10
  21. 2024Are Large Language Models Really Bias-Free? Jailbreak Prompts for Assessing Adversarial Robustness to Bias Elicitationcited by 9
  22. 2021Deciphering Implicit Hate: Evaluating Automated Detection Algorithms for Multimodal Hatecited by 8
  23. 2022XMD: An End-to-End Framework for Interactive Explanation-Based Debugging of NLP Modelscited by 8
  24. 2021Character-level HyperNetworks for Hate Speech Detectioncited by 7
  25. 2022Data-Efficient Strategies for Expanding Hate Speech Detection into Under-Resourced Languagescited by 7
  26. 2023Generative AI for Hate Speech Detection: Evaluation and Findingscited by 7
  27. 2024ToXCL: A Unified Framework for Toxic Speech Detection and Explanationcited by 7
  28. 2024Hate Speech Detection with Generalizable Target-aware Fairnesscited by 7
  29. 2022ER-Test: Evaluating Explanation Regularization Methods for Language Modelscited by 6
  30. 2022Assessing Out-of-Domain Language Model Performance from Few Examplescited by 6
  31. 2022Towards Procedural Fairness: Uncovering Biases in How a Toxic Language Classifier Uses Sentiment Informationcited by 6
  32. 2022Power of Explanations: Towards automatic debiasing in hate speech detectioncited by 5
  33. 2026Self-Explaining Hate Speech Detection with Moral Rationalescited by 5
  34. 2021On a Benefit of Mask Language Modeling: Robustness to Simplicity Biascited by 4
  35. 2023REFER: An End-to-end Rationale Extraction Framework for Explanation Regularizationcited by 4
  36. 2024Exploring Boundaries and Intensities in Offensive and Hate Speech: Unveiling the Complex Spectrum of Social Media Discoursecited by 4
  37. 2025Characterizing Selective Refusal Bias in Large Language Modelscited by 4
  38. 2022Dynamically Refined Regularization for Improving Cross-corpora Hate Speech Detectioncited by 3

50 most cited of 84 All 84 (JSON)