mlsanityData to 7 Oct 2026

Polyjuice: Generating Counterfactuals for Explaining, Evaluating, and Improving Models

arXiv 2101.00288 ↗First posted 1 Jan 2021cs.CLcited by 219

Not on the map. Why?

Abstract (arXiv)

While counterfactual examples are useful for analysis and training of NLP models, current generation methods either rely on manual labor to create very few counterfactuals, or only instantiate limited types of perturbations such as paraphrases or word substitutions. We present Polyjuice, a general-purpose counterfactual generator that allows for control over perturbation types and locations, trained by finetuning GPT-2 on multiple datasets of paired sentences. We show that Polyjuice produces diverse sets of realistic counterfactuals, which in turn are useful in various distinct applications: improving training and evaluation on three different tasks (with around 70% less annotation effort than manual generation), augmenting state-of-the-art explanation techniques, and supporting systematic counterfactual error analysis by revealing behaviors easily missed by human experts.

Key facts. Cited by 219 papers in the map (data through 7 Oct 2026). Builds on 39 of its 66 references that are in the map. First posted on arXiv 1 Jan 2021.

Builds on

39 of 66 references.

References · 39

  1. 2019Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformercited by 14,491
  2. 2019RoBERTa: A Robustly Optimized BERT Pretraining Approachcited by 13,266
  3. 2019Sentence-BERT: Sentence Embeddings using Siamese BERT-Networkscited by 9,049
  4. 2017A Unified Approach to Interpreting Model Predictionscited by 6,488
  5. 2016"Why Should I Trust You?": Explaining the Predictions of Any Classifiercited by 5,040
  6. 2018GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understandingcited by 4,854
  7. 2016SQuAD: 100,000+ Questions for Machine Comprehension of Textcited by 4,115
  8. 2015A large annotated corpus for learning natural language inferencecited by 1,837
  9. 2017Explanation in Artificial Intelligence: Insights from the Social Sciencescited by 1,201
  10. 2020Beyond Accuracy: Behavioral Testing of NLP models with CheckListcited by 979
  11. 2019Right for the Wrong Reasons: Diagnosing Syntactic Heuristics in Natural Language Inferencecited by 864
  12. 2019Plug and Play Language Models: A Simple Approach to Controlled Text Generationcited by 793
27 more
  1. 2019CTRL: A Conditional Transformer Language Model for Controllable Generationcited by 754
  2. 2018Annotation Artifacts in Natural Language Inference Datacited by 656
  3. 2019Attention is not not Explanationcited by 629
  4. 2018Texygen: A Benchmarking Platform for Text Generation Modelscited by 484
  5. 2019Learning the Difference that Makes a Difference with Counterfactually-Augmented Datacited by 428
  6. 2020Does the Whole Exceed its Parts? The Effect of AI Explanations on Complementary Team Performancecited by 396
  7. 2019PAWS: Paraphrase Adversaries from Word Scramblingcited by 395
  8. 2018Adversarial Example Generation with Syntactically Controlled Paraphrase Networkscited by 282
  9. 2018Stress Test Evaluation for Natural Language Inferencecited by 211
  10. 2020Evaluating Explainable AI: Which Algorithmic Explanations Help Users Predict Model Behavior?cited by 189
  11. 2018Counterfactual Fairness in Text Classification through Robustnesscited by 150
  12. 2020Contextualized Perturbation for Textual Adversarial Attackcited by 136
  13. 2018Breaking NLI Systems with Sentences that Require Simple Lexical Inferencescited by 134
  14. 2020Enabling Language Models to Fill in the Blankscited by 121
  15. 2019Learning to Deceive with Attention-Based Explanationscited by 111
  16. 2020Explaining NLP Models via Minimal Contrastive Editing (MiCE)cited by 105
  17. 2019Counterfactual Story Reasoning and Generationcited by 87
  18. 2016Yelp Dataset Challenge: Review Rating Predictioncited by 84
  19. 2020Politeness Transfer: A Tag and Generate Approachcited by 78
  20. 2020Generate Your Counterfactuals: Towards Controlled Counterfactual Generation for Textcited by 75
  21. 2020Learning What Makes a Difference from Counterfactual Examples and Gradient Supervisioncited by 66
  22. 2020Generating Fluent Adversarial Examples for Natural Languagescited by 49
  23. 2019Probing What Different NLP Tasks Teach Machines about Function Word Comprehensioncited by 44
  24. 2020How Effective is Task-Agnostic Data Augmentation for Pretrained Transformers?cited by 41
  25. 2020Counterfactually-Augmented SNLI Training Data Does Not Yield Better Generalization Than Unaugmented Datacited by 27
  26. 2019Slice-based Learning: A Programming Model for Residual Learning in Critical Data Slicescited by 23
  27. 2019SemEval-2013 Task 2: Sentiment Analysis in Twittercited by 22

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 20; 2022: 47; 2023: 57; 2024: 47; 2025: 24; 2026: 24.

219 papers.

Sort:

By area:

  1. 2022Red Teaming Language Models with Language Modelscited by 983
  2. 2022Discovering Language Model Behaviors with Model-Written Evaluationscited by 805
  3. 2021Recent Advances in Natural Language Processing via Large Pre-Trained Language Models: A Surveycited by 457
  4. 2023Explainability for Large Language Models: A Surveycited by 331
  5. 2021AI Chains: Transparent and Controllable Human-AI Interaction by Chaining Large Language Model Promptscited by 265
  6. 2023Clever Hans or Neural Theory of Mind? Stress Testing Social Reasoning in Large Language Modelscited by 161
  7. 2021Post-hoc Interpretability for Neural NLP: A Surveycited by 155
  8. 2022Rich Knowledge Sources Bring Complex Knowledge Conflicts: Recalibrating Models to Reflect Conflicting Evidencecited by 142
  9. 2021Benchmarking and Survey of Explanation Methods for Black Box Modelscited by 136
  10. 2022Towards Faithful Model Explanation in NLP: A Surveycited by 129
  11. 2021Teach Me to Explain: A Review of Datasets for Explainable Natural Language Processingcited by 126
  12. 2021Measure and Improve Robustness in NLP Models: A Surveycited by 107
38 more
  1. 2023Faithfulness Tests for Natural Language Explanationscited by 107
  2. 2024Are self-explanations from Large Language Models faithful?cited by 101
  3. 2023Revisiting Out-of-distribution Robustness in NLP: Benchmark, Analysis, and LLMs Evaluationscited by 96
  4. 2023Do Models Explain Themselves? Counterfactual Simulatability of Natural Language Explanationscited by 85
  5. 2021Counterfactual Invariance to Spurious Correlations: Why and How to Pass Stress Testscited by 77
  6. 2021Contrastive Explanations for Model Interpretabilitycited by 75
  7. 2023EvalLM: Interactive Evaluation of Large Language Model Prompts on User-Defined Criteriacited by 73
  8. 2021NL-Augmenter: A Framework for Task-Sensitive Natural Language Augmentationcited by 72
  9. 2022GLUE-X: Evaluating Natural Language Understanding Models from an Out-of-distribution Generalization Perspectivecited by 72
  10. 2025Interactive Debugging and Steering of Multi-Agent AI Systemscited by 68
  11. 2021Tailor: Generating and Perturbing Text with Semantic Controlscited by 67
  12. 2023From Language Modeling to Instruction Following: Understanding the Behavior Shift in LLMs after Instruction Tuningcited by 56
  13. 2023SocialCounterfactuals: Probing and Mitigating Intersectional Social Biases in Vision-Language Models with Counterfactual Examplescited by 53
  14. 2022Leveraging Explanations in Interactive Machine Learning: An Overviewcited by 52
  15. 2022Perturbation Augmentation for Fairer NLPcited by 51
  16. 2023TIBET: Identifying and Evaluating Biases in Text-to-Image Generative Modelscited by 51
  17. 2023Towards LLM-guided Causal Explainability for Black-box Text Classifierscited by 44
  18. 2021An Investigation of the (In)effectiveness of Counterfactually Augmented Datacited by 42
  19. 2023COCO-Counterfactuals: Automatically Constructed Counterfactual Examples for Image-Text Pairscited by 39
  20. 2022DoCoGen: Domain Counterfactual Generation for Low Resource Domain Adaptationcited by 36
  21. 2023Interpreting Pretrained Language Models via Concept Bottleneckscited by 33
  22. 2023Using Captum to Explain Generative Language Modelscited by 33
  23. 2021Local Interpretations for Explainable Natural Language Processing: A Surveycited by 31
  24. 2021Models in the Loop: Aiding Crowdworkers with Generative Annotation Assistantscited by 31
  25. 2023On Evaluating and Mitigating Gender Biases in Multilingual Settingscited by 31
  26. 2023Quantifying Uncertainty in Natural Language Explanations of Large Language Modelscited by 31
  27. 2024LLMs for Generating and Evaluating Counterfactuals: A Comprehensive Studycited by 30
  28. 2022Causal Proxy Models for Concept-Based Model Explanationscited by 29
  29. 2022DISCO: Distilling Counterfactuals with Large Language Modelscited by 29
  30. 2021Retrieval-guided Counterfactual Generation for QAcited by 28
  31. 2023LLMs as Workers in Human-Computational Algorithms? Replicating Crowdsourcing Pipelines with LLMscited by 27
  32. 2025Is Conversational XAI All You Need? Human-AI Decision Making With a Conversational XAI Assistantcited by 26
  33. 2021Counterfactual Explanations for Models of Codecited by 24
  34. 2023CREST: A Joint Framework for Rationalization and Counterfactual Text Generationcited by 24
  35. 2024On Behalf of the Stakeholders: Trends in NLP Model Interpretability in the Era of LLMscited by 24
  36. 2022CORE: A Retrieve-then-Edit Framework for Counterfactual Data Generationcited by 23
  37. 2022NeuroCounterfactuals: Beyond Minimal-Edit Counterfactuals for Richer Data Augmentationcited by 23
  38. 2024XAI meets LLMs: A Survey of the Relation between Explainable AI and Large Language Modelscited by 23

50 most cited of 219 All 219 (JSON)