mlsanityData to 14 Oct 2026

Evaluating Gender Bias in Natural Language Inference

arXiv 2105.05541 ↗First posted 12 May 2021cs.CL cs.LGcited by 27

Not on the map. Why?

Abstract (arXiv)

Gender-bias stereotypes have recently raised significant ethical concerns in natural language processing. However, progress in detection and evaluation of gender bias in natural language understanding through inference is limited and requires further investigation. In this work, we propose an evaluation methodology to measure these biases by constructing a challenge task that involves pairing gender-neutral premises against a gender-specific hypothesis. We use our challenge task to investigate state-of-the-art NLI models on the presence of gender stereotypes using occupations. Our findings suggest that three models (BERT, RoBERTa, BART) trained on MNLI and SNLI datasets are significantly prone to gender-induced prediction errors. We also find that debiasing techniques such as augmenting the training dataset to ensure a gender-balanced dataset can help reduce such bias in certain cases.

Key facts. Cited by 27 papers in the map (data through 14 Oct 2026). Builds on 29 of its 35 references that are in the map. First posted on arXiv 12 May 2021.

Builds on

29 of 35 references.

References · 29

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
  2. 2019RoBERTa: A Robustly Optimized BERT Pretraining Approachcited by 13,372
  3. 2019BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehensioncited by 5,893
  4. 2016SQuAD: 100,000+ Questions for Machine Comprehension of Textcited by 4,148
  5. 2017A Broad-Coverage Challenge Corpus for Sentence Understanding through Inferencecited by 2,682
  6. 2015A large annotated corpus for learning natural language inferencecited by 1,847
  7. 2016Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddingscited by 1,312
  8. 2016Semantics derived automatically from language corpora contain human-like biasescited by 901
  9. 2019Right for the Wrong Reasons: Diagnosing Syntactic Heuristics in Natural Language Inferencecited by 876
  10. 2019Adversarial NLI: A New Benchmark for Natural Language Understandingcited by 807
  11. 2018Gender Bias in Coreference Resolution: Evaluation and Debiasing Methodscited by 659
  12. 2017Adversarial Examples for Evaluating Reading Comprehension Systemscited by 571
17 more
  1. 2018Gender Bias in Coreference Resolutioncited by 442
  2. 2019Benchmarking Zero-shot Text Classification: Datasets, Evaluation and Entailment Approachcited by 308
  3. 2019Don't Take the Easy Way Out: Ensemble Based Methods for Avoiding Known Dataset Biasescited by 280
  4. 2019Lipstick on a Pig: Debiasing Methods Cover up Systematic Gender Biases in Word Embeddings But do not Remove Themcited by 268
  5. 2018Learning Gender-Neutral Word Embeddingscited by 169
  6. 2019Good-Enough Compositional Data Augmentationcited by 158
  7. 2018Reducing Gender Bias in Abusive Language Detectioncited by 152
  8. 2019Queens are Powerful too: Mitigating Gender Bias in Dialogue Generationcited by 152
  9. 2018Breaking NLI Systems with Sentences that Require Simple Lexical Inferencescited by 134
  10. 2017Dataset Augmentation in Feature Spacecited by 124
  11. 2019Unlearn Dataset Bias in Natural Language Inference by Fitting the Residualcited by 119
  12. 2018Understanding the Origins of Bias in Word Embeddingscited by 91
  13. 2017Ethical Challenges in Data-Driven Dialogue Systemscited by 86
  14. 2020Syntactic Data Augmentation Increases Robustness to Inference Heuristicscited by 86
  15. 2020Towards Robustifying NLI Models Against Lexical Dataset Biasescited by 34
  16. 2020Evaluating Models' Local Decision Boundaries via Contrast Setscited by 24
  17. 2019Towards Understanding Gender Bias in Relation Extractioncited by 23

Built on by

Cited by year

Citing papers per year: 2021: 2; 2022: 13; 2023: 10; 2024: 1; 2025: 1.

27 papers.

Sort:

By area:

  1. 2021BBQ: A Hand-Built Bias Benchmark for Question Answeringcited by 701
  2. 2022Theories of "Gender" in NLP Bias Researchcited by 62
  3. 2022Perturbation Augmentation for Fairer NLPcited by 52
  4. 2022Fewer Errors, but More Stereotypes? The Effect of Model Size on Gender Biascited by 33
  5. 2024Can LLMs Recognize Toxicity? A Structured Investigation Framework and Toxicity Metriccited by 23
  6. 2023Evaluating the Robustness of Discrete Promptscited by 18
  7. 2023Evaluating Gender Bias of Pre-trained Language Models in Natural Language Inference by Considering All Labelscited by 16
  8. 2023What Makes it Ok to Set a Fire? Iterative Self-distillation of Contexts and Rationales for Disambiguating Defeasible Social and Moral Situationscited by 15
  9. 2022How Gender Debiasing Affects Internal Model Representations, and Why It Matterscited by 12
  10. 2022JASMINE: Arabic GPT Models for Few-Shot Learningcited by 11
  11. 2022How sensitive are translation systems to extra contexts? Mitigating gender bias in Neural Machine Translation models through relevant contextscited by 10
  12. 2022On Measuring Social Biases in Prompt-Based Multi-Task Learningcited by 10
15 more
  1. 2023Target-Agnostic Gender-Aware Contrastive Learning for Mitigating Bias in Multilingual Machine Translationcited by 10
  2. 2023Prompt Valuation Based on Shapley Valuescited by 10
  3. 2023Mitigating Bias for Question Answering Models by Tracking Bias Influencecited by 8
  4. 2022FAST: Improving Controllability for Text Generation with Feedback Aware Self-Trainingcited by 7
  5. 2022Data Representativeness in Accessibility Datasets: A Meta-Analysiscited by 6
  6. 2022LawngNLI: A Long-Premise Benchmark for In-Domain Generalization from Short to Long Contexts and for Implication-Based Retrievalcited by 6
  7. 2023Tailoring with Targeted Precision: Edit-Based Agents for Open-Domain Procedure Customizationcited by 5
  8. 2022Social Biases in Automatic Evaluation Metrics for NLGcited by 3
  9. 2023Disentangling Societal Inequality from Model Biases: Gender Inequality in Divorce Court Proceedingscited by 3
  10. 2023Learning to Predict Concept Ordering for Common Sense Generationcited by 3
  11. 2022Memory-assisted prompt editing to improve GPT-3 after deploymentcited by 2
  12. 2022Subspace Representations for Soft Set Operations and Sentence Similaritiescited by 2
  13. 2025Measuring Gender Bias in Job Title Matching for Grammatical Gender Languagescited by 2
  14. 2021Logically at Factify 2022: Multimodal Fact Verificationcited by 0
  15. 2023AMR4NLI: Interpretable and robust NLI measures from semantic graphscited by 0