mlsanityData to 7 Oct 2026

Identifying and Mitigating Spurious Correlations for Improving Robustness in NLP Models

arXiv 2110.07736 ↗First posted 14 Oct 2021cs.CLcited by 76

Not on the map. Why?

Abstract (arXiv)

Recently, NLP models have achieved remarkable progress across a variety of tasks; however, they have also been criticized for being not robust. Many robustness problems can be attributed to models exploiting spurious correlations, or shortcuts between the training data and the task labels. Most existing work identifies a limited set of task-specific shortcuts via human priors or error analyses, which requires extensive expertise and efforts. In this paper, we aim to automatically identify such spurious correlations in NLP models at scale. We first leverage existing interpretability methods to extract tokens that significantly affect model's decision process from the input text. We then distinguish "genuine" tokens and "spurious" tokens by analyzing model predictions across multiple corpora and further verify them through knowledge-aware perturbations. We show that our proposed method can effectively and efficiently identify a scalable set of "shortcuts", and mitigating these leads to more robust models in multiple applications.

Key facts. Cited by 76 papers in the map (data through 7 Oct 2026). Builds on 37 of its 48 references that are in the map. First posted on arXiv 14 Oct 2021.

Builds on

37 of 48 references.

References · 37

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  2. 2016"Why Should I Trust You?": Explaining the Predictions of Any Classifiercited by 5,040
  3. 2017Axiomatic Attribution for Deep Networkscited by 3,350
  4. 2020Shortcut Learning in Deep Neural Networkscited by 1,928
  5. 2019What Does BERT Look At? An Analysis of BERT's Attentioncited by 994
  6. 2020Beyond Accuracy: Behavioral Testing of NLP models with CheckListcited by 979
  7. 2018Robustness May Be at Odds with Accuracycited by 873
  8. 2019Right for the Wrong Reasons: Diagnosing Syntactic Heuristics in Natural Language Inferencecited by 864
  9. 2017Adversarial Examples for Evaluating Reading Comprehension Systemscited by 568
  10. 2016Ups and Downs: Modeling the Visual Evolution of Fashion Trends with One-Class Collaborative Filteringcited by 517
  11. 2018Generating Natural Language Adversarial Examplescited by 422
  12. 2020An Investigation of Why Overparameterization Exacerbates Spurious Correlationscited by 310
25 more
  1. 2019Revealing the Dark Secrets of BERTcited by 300
  2. 2020Pretrained Transformers Improve Out-of-Distribution Robustnesscited by 293
  3. 2019Don't Take the Easy Way Out: Ensemble Based Methods for Avoiding Known Dataset Biasescited by 279
  4. 2020Adversarial Filters of Dataset Biasescited by 159
  5. 2020WT5?! Training Text-to-Text Models to Explain their Predictionscited by 148
  6. 2016Counter-fitting Word Vectors to Linguistic Constraintscited by 138
  7. 2020An Empirical Study on Robustness to Spurious Correlations using Pre-trained Language Modelscited by 133
  8. 2020Self-Attention Attribution: Interpreting Information Interactions Inside Transformercited by 124
  9. 2019Unlearn Dataset Bias in Natural Language Inference by Fitting the Residualcited by 119
  10. 2019Learning to Deceive with Attention-Based Explanationscited by 111
  11. 2016Yelp Dataset Challenge: Review Rating Predictioncited by 84
  12. 2019Fairwashing: the risk of rationalizationcited by 81
  13. 2021Towards Interpreting and Mitigating Shortcut Learning Behavior of NLU Modelscited by 78
  14. 2021Contrastive Explanations for Model Interpretabilitycited by 75
  15. 2020Robustness to Spurious Correlations in Text Classification via Automatically Generated Counterfactualscited by 67
  16. 2020Robustness to Spurious Correlations via Human Annotationscited by 62
  17. 2020Identifying Spurious Correlations for Robust Text Classificationcited by 59
  18. 2020Generating Hierarchical Explanations on Text Classification via Feature Interaction Detectioncited by 57
  19. 2020Removing Spurious Features can Hurt Accuracy and Affect Groups Disproportionatelycited by 52
  20. 2020Automatic Shortcut Removal for Self-Supervised Representation Learningcited by 43
  21. 2020Learning to Model and Ignore Dataset Bias with Mixed Capacity Ensemblescited by 43
  22. 2020Evaluating Robustness to Input Perturbations for Neural Machine Translationcited by 36
  23. 2020Towards Robustifying NLI Models Against Lexical Dataset Biasescited by 34
  24. 2021HiddenCut: Simple Data Augmentation for Natural Language Understanding with Better Generalizationcited by 31
  25. 2019Adversarial Domain Adaptation for Machine Reading Comprehensioncited by 21

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 1; 2022: 10; 2023: 26; 2024: 17; 2025: 12; 2026: 10.

76 papers.

Sort:

By area:

  1. 2023Inverse Scaling: When Bigger Isn't Bettercited by 205
  2. 2021Measure and Improve Robustness in NLP Models: A Surveycited by 107
  3. 2023Revisiting Out-of-distribution Robustness in NLP: Benchmark, Analysis, and LLMs Evaluationscited by 96
  4. 2023Language Model Behavior: A Comprehensive Surveycited by 92
  5. 2022Shortcut Learning of Large Language Models in Natural Language Understandingcited by 83
  6. 2023Large Language Models Can be Lazy Learners: Analyze Shortcuts in In-Context Learningcited by 69
  7. 2022A Close Look into the Calibration of Pre-trained Language Modelscited by 57
  8. 2022A Comprehensive Study of Image Classification Model Sensitivity to Foregrounds, Backgrounds, and Visual Attributescited by 56
  9. 2023Explore Spurious Correlations at the Concept Level in Language Models for Text Classificationcited by 48
  10. 2023RobuT: A Systematic Study of Table QA Robustness Against Human-Annotated Adversarial Perturbationscited by 41
  11. 2024InjecGuard: Benchmarking and Mitigating Over-defense in Prompt Injection Guardrail Modelscited by 36
  12. 2024MM-SpuBench: Towards Better Understanding of Spurious Biases in Multimodal LLMscited by 33
38 more
  1. 2023Towards Trustworthy and Aligned Machine Learning: A Data-centric Survey with Causality Perspectivescited by 30
  2. 2022A Rationale-Centric Framework for Human-in-the-loop Machine Learningcited by 27
  3. 2023Impact of Co-occurrence on Factual Knowledge of Large Language Modelscited by 19
  4. 2024Disperse-Then-Merge: Pushing the Limits of Instruction Tuning via Alignment Tax Reductioncited by 18
  5. 2025Morae: Proactively Pausing UI Agents for User Choicescited by 18
  6. 2022Does Your Model Classify Entities Reasonably? Diagnosing and Mitigating Spurious Correlations in Entity Typingcited by 16
  7. 2022Finding Dataset Shortcuts with Grammar Inductioncited by 15
  8. 2023Shift-Robust Molecular Relational Learning with Causal Substructurecited by 15
  9. 2023Evaluating Neural Language Models as Cognitive Models of Language Acquisitioncited by 15
  10. 2023Understanding and Mitigating Spurious Correlations in Text Classification with Neighborhood Analysiscited by 11
  11. 2024What makes a good metric? Evaluating automatic metrics for text-to-image consistencycited by 11
  12. 2023Guiding LLM to Fool Itself: Automatically Manipulating Machine Reading Comprehension Shortcut Triggerscited by 10
  13. 2024Navigating the Shortcut Maze: A Comprehensive Analysis of Shortcut Learning in Text Classification by Language Modelscited by 10
  14. 2023Prototype-Based Interpretability for Legal Citation Predictioncited by 9
  15. 2023Why Linguistics Will Thrive in the 21st Century: A Reply to Piantadosi (2023)cited by 9
  16. 2024Toward the Evaluation of Large Language Models Considering Score Variance across Instruction Templatescited by 7
  17. 2025Bias Is a Subspace, Not a Coordinate: A Geometric Rethinking of Post-hoc Debiasing in Vision-Language Modelscited by 7
  18. 2023Spurious Correlations and Where to Find Themcited by 6
  19. 2023Predict the Future from the Past? On the Temporal Data Distribution Shift in Financial Sentiment Classificationscited by 6
  20. 2024Improving LLM Group Fairness on Tabular Data via In-Context Learningcited by 6
  21. 2023Controlling Learned Effects to Reduce Spurious Correlations in Text Classifierscited by 5
  22. 2024Focus On This, Not That! Steering LLMs with Adaptive Feature Specificationcited by 5
  23. 2024RAZOR: Sharpening Knowledge by Cutting Bias with Unsupervised Text Rewritingcited by 5
  24. 2025Short-circuiting Shortcuts: Mechanistic Investigation of Shortcuts in Text Classificationcited by 5
  25. 2025Evaluating and Improving Robustness in Large Language Models: A Survey and Future Directionscited by 5
  26. 2022A Survey on Measuring and Mitigating Reasoning Shortcuts in Machine Reading Comprehensioncited by 4
  27. 2022Less is Better: Recovering Intended-Feature Subspace to Robustify NLU Modelscited by 4
  28. 2024Learning Shortcuts: On the Misleading Promise of NLU in Language Modelscited by 4
  29. 2022On Text-based Personality Computing: Challenges and Future Directionscited by 3
  30. 2023Consistent Text Categorization using Data Augmentation in e-Commercecited by 3
  31. 2023Did the Models Understand Documents? Benchmarking Models for Language Understanding in Document-Level Relation Extractioncited by 3
  32. 2024Towards Operationalizing Right to Data Protectioncited by 3
  33. 2026Chunky Post-Training: Data Driven Failures of Generalizationcited by 3
  34. 2023GLS-CSC: A Simple but Effective Strategy to Mitigate Chinese STM Models' Over-Reliance on Superficial Cluecited by 2
  35. 2023Mitigating Shortcuts in Language Models with Soft Label Encodingcited by 2
  36. 2023Discovering Highly Influential Shortcut Reasoning: An Automated Template-Free Approachcited by 2
  37. 2024Specification Overfitting in Artificial Intelligencecited by 2
  38. 2024Towards Data-Centric Automatic R&Dcited by 2

50 most cited of 76 All 76 (JSON)