mlsanityData to 7 Oct 2026

Measuring and Reducing Gendered Correlations in Pre-trained Models

arXiv 2010.06032 ↗First posted 12 Oct 2020cs.CLcited by 222

Not on the map. Why?

Abstract (arXiv)

Pre-trained models have revolutionized natural language understanding. However, researchers have found they can encode artifacts undesired in many applications, such as professions correlating with one gender more than another. We explore such gendered correlations as a case study for how to address unintended correlations in pre-trained models. We define metrics and reveal that it is possible for models with similar accuracy to encode correlations at very different rates. We show how measured correlations can be reduced with general-purpose techniques, and highlight the trade offs different strategies have. With these results, we make recommendations for training robust models: (1) carefully evaluate unintended correlations, (2) be mindful of seemingly innocuous configuration differences, and (3) focus on general mitigations.

Key facts. Cited by 222 papers in the map (data through 7 Oct 2026). First posted on arXiv 12 Oct 2020.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 17; 2022: 37; 2023: 49; 2024: 63; 2025: 45; 2026: 11.

222 papers.

Sort:

By area:

  1. 2021GLaM: Efficient Scaling of Language Models with Mixture-of-Expertscited by 890
  2. 2023Bias and Fairness in Large Language Models: A Surveycited by 626
  3. 2022Challenges and Strategies in Cross-Cultural NLPcited by 216
  4. 2022"I'm sorry to hear that": Finding New Biases in Language Models with a Holistic Descriptor Datasetcited by 160
  5. 2021RedditBias: A Real-World Resource for Bias Evaluation and Debiasing of Conversational Language Modelscited by 149
  6. 2023Auditing large language models: a three-layered approachcited by 135
  7. 2022Pretrained Language Models for Text Generation: A Surveycited by 121
  8. 2021Sustainable Modular Debiasing of Language Modelscited by 98
  9. 2021An Empirical Survey of the Effectiveness of Debiasing Techniques for Pre-trained Language Modelscited by 96
  10. 2023A Survey on Fairness in Large Language Modelscited by 91
  11. 2024Large Language Models are Geographically Biasedcited by 91
  12. 2021The MultiBERTs: BERT Reproductions for Robustness Analysiscited by 89
38 more
  1. 2024A Survey on Multilingual Large Language Models: Corpora, Alignment, and Biascited by 82
  2. 2024Bias in Large Language Models: Origin, Evaluation, and Mitigationcited by 77
  3. 2023Word Embeddings Are Steers for Language Modelscited by 70
  4. 2023Fast Model Debias with Machine Unlearningcited by 70
  5. 2024Navigating the Risks: A Survey of Security, Privacy, and Ethics Threats in LLM-Based Agentscited by 63
  6. 2024Self-Debiasing Large Language Models: Zero-Shot Recognition and Reduction of Stereotypescited by 58
  7. 2024Fairness in Large Language Models: A Taxonomic Surveycited by 57
  8. 2023TrustGPT: A Benchmark for Trustworthy and Responsible Large Language Modelscited by 56
  9. 2022Re-contextualizing Fairness in NLP: The Case of Indiacited by 55
  10. 2021Worst of Both Worlds: Biases Compound in Pre-trained Vision-and-Language Modelscited by 52
  11. 2022Perturbation Augmentation for Fairer NLPcited by 51
  12. 2024Evaluating Gender Bias in Large Language Models via Chain-of-Thought Promptingcited by 49
  13. 2022Debiasing isn't enough! -- On the Effectiveness of Debiasing MLMs and their Social Biases in Downstream Taskscited by 48
  14. 2022BERTScore is Unfair: On Social Bias in Language Model-Based Metrics for Text Generationcited by 48
  15. 2024Thinking Fair and Slow: On the Efficacy of Structured Prompts for Debiasing Language Modelscited by 44
  16. 2021Your fairness may vary: Pretrained language model fairness in toxic text classificationcited by 38
  17. 2025FairSteer: Inference Time Debiasing for LLMs with Dynamic Activation Steeringcited by 37
  18. 2021Improving Gender Fairness of Pre-Trained Language Models without Catastrophic Forgettingcited by 34
  19. 2022Validating Large Language Models with ReLMcited by 33
  20. 2024Graph Machine Learning in the Era of Large Language Models (LLMs)cited by 33
  21. 2022MABEL: Attenuating Gender Bias using Textual Entailment Datacited by 31
  22. 2023On Evaluating and Mitigating Gender Biases in Multilingual Settingscited by 31
  23. 2022Choose Your Lenses: Flaws in Gender Bias Evaluationcited by 29
  24. 2021Retrieval-guided Counterfactual Generation for QAcited by 28
  25. 2022Counterfactually Measuring and Eliminating Social Bias in Vision-Language Pre-training Modelscited by 28
  26. 2022Causally Testing Gender Bias in LLMs: A Case Study on Occupational Biascited by 28
  27. 2023Language Models Get a Gender Makeover: Mitigating Gender Bias with Few-Shot Data Interventionscited by 28
  28. 2023Sociodemographic Bias in Language Models: A Survey and Forward Pathcited by 27
  29. 2022Toxicity in Multilingual Machine Translation at Scalecited by 26
  30. 2023Interpretable Unified Language Checkingcited by 26
  31. 2023Survey of Social Bias in Vision-Language Modelscited by 26
  32. 2023Uncovering and Quantifying Social Biases in Code Generationcited by 25
  33. 2022A Robust Bias Mitigation Procedure Based on the Stereotype Content Modelcited by 24
  34. 2024Causality for Large Language Modelscited by 24
  35. 2021Hi, my name is Martha: Using names to measure and mitigate bias in generative dialogue modelscited by 23
  36. 2022Unit Testing for Concepts in Neural Networkscited by 23
  37. 2022CORE: A Retrieve-then-Edit Framework for Counterfactual Data Generationcited by 23
  38. 2023Multilingual Holistic Bias: Extending Descriptors and Patterns to Unveil Demographic Biases in Languages at Scalecited by 23

50 most cited of 222 All 222 (JSON)