mlsanityData to 7 Oct 2026

Automatically Neutralizing Subjective Bias in Text

arXiv 1911.09709 ↗First posted 21 Nov 2019cs.CL cs.AIcited by 105

Not on the map. Why?

Abstract (arXiv)

Texts like news, encyclopedias, and some social media strive for objectivity. Yet bias in the form of inappropriate subjectivity - introducing attitudes via framing, presupposing truth, and casting doubt - remains ubiquitous. This kind of bias erodes our collective trust and fuels social conflict. To address this issue, we introduce a novel testbed for natural language generation: automatically bringing inappropriately subjective text into a neutral point of view ("neutralizing" biased text). We also offer the first parallel corpus of biased language. The corpus contains 180,000 sentence pairs and originates from Wikipedia edits that removed various framings, presuppositions, and attitudes from biased sentences. Last, we propose two strong encoder-decoder baselines for the task. A straightforward yet opaque CONCURRENT system uses a BERT encoder to identify subjective words as part of the generation process. An interpretable and controllable MODULAR algorithm separates these steps, using (1) a BERT-based classifier to identify problematic words and (2) a novel join embedding through which the classifier can edit the hidden states of the encoder. Large-scale human evaluation across four domains (encyclopedias, news headlines, books, and political speeches) suggests that these algorithms are a first step towards the automatic identification and reduction of bias.

Key facts. Cited by 105 papers in the map (data through 7 Oct 2026). First posted on arXiv 21 Nov 2019.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 18; 2022: 18; 2023: 29; 2024: 25; 2025: 10; 2026: 5.

105 papers.

Sort:

By area:

  1. 2023Bias and Fairness in Large Language Models: A Surveycited by 626
  2. 2021Societal Biases in Language Generation: Progress and Challengescited by 201
  3. 2022A Survey of Controllable Text Generation using Transformer-based Pre-trained Language Modelscited by 128
  4. 2024Symbolic Learning Enables Self-Evolving Agentscited by 104
  5. 2023Navigating the Grey Area: How Expressions of Uncertainty and Overconfidence Affect Language Modelscited by 91
  6. 2023Controlled Text Generation with Natural Language Instructionscited by 87
  7. 2022PEER: A Collaborative Language Modelcited by 86
  8. 2022Language Generation Models Can Cause Harm: So What Can We Do About It? An Actionable Surveycited by 80
  9. 2021Towards Facilitating Empathic Conversations in Online Mental Health Support: A Reinforcement Learning Approachcited by 75
  10. 2023CoEdIT: Text Editing by Task-Specific Instruction Tuningcited by 67
  11. 2023RewriteLM: An Instruction-Tuned Large Language Model for Text Rewritingcited by 67
  12. 2023Do LLMs Understand Social Knowledge? Evaluating the Sociability of Large Language Models with SocKET Benchmarkcited by 62
38 more
  1. 2022LENS: A Learnable Evaluation Metric for Text Simplificationcited by 54
  2. 2024Can AI writing be salvaged? Mitigating Idiosyncrasies and Improving Human-AI Alignment in the Writing Process through Editscited by 49
  3. 2023Are Personalized Stochastic Parrots More Dangerous? Evaluating Persona Biases in Dialogue Systemscited by 37
  4. 2023Beyond the Chat: Executable and Verifiable Text-Editing with LLMscited by 35
  5. 2023Multi-Task Instruction Tuning of LLaMa for Specific Scenarios: A Preliminary Study on Writing Assistancecited by 33
  6. 2021Civil Rephrases Of Toxic Texts With Self-Supervised Transformerscited by 32
  7. 2022Inducing Positive Perspectives with Text Reframingcited by 31
  8. 2022EditEval: An Instruction-Based Benchmark for Text Improvementscited by 30
  9. 2023Learning Interpretable Style Embeddings via Prompting LLMscited by 29
  10. 2022NEWTS: A Corpus for News Topic-Focused Summarizationcited by 27
  11. 2022arXivEdits: Understanding the Human Revision Process in Scientific Writingcited by 26
  12. 2021Detect and Perturb: Neutral Rewriting of Biased and Sensitive Text via Gradient-based Decodingcited by 23
  13. 2023NBIAS: A Natural Language Processing Framework for Bias Identification in Textcited by 23
  14. 2022Gender Biases and Where to Find Them: Exploring Gender Bias in Pre-Trained Transformer-based Language Models Using Movement Pruningcited by 20
  15. 2024REVOLVE: Optimizing AI Systems by Tracking Response Evolution in Textual Optimizationcited by 20
  16. 2021A Review of Human Evaluation for Style Transfercited by 19
  17. 2024Fairness and Bias in Multimodal AI: A Surveycited by 19
  18. 2022Proactively Reducing the Hate Intensity of Online Posts via Hate Speech Normalizationcited by 17
  19. 2023Introducing MBIB -- the first Media Bias Identification Benchmark Task and Dataset Collectioncited by 17
  20. 2022A Domain-adaptive Pre-training Approach for Language Bias Detection in Newscited by 16
  21. 2021Neural semi-Markov CRF for Monolingual Word Alignmentcited by 15
  22. 2023Exploiting Biased Models to De-bias Text: A Gender-Fair Rewriting Modelcited by 14
  23. 2021Methods for Detoxification of Texts for the Russian Languagecited by 13
  24. 2023A Call for Standardization and Validation of Text Style Transfer Evaluationcited by 13
  25. 2021Style Pooling: Automatic Text Style Obfuscation for Improved Classification Fairnesscited by 12
  26. 2021Transductive Learning for Unsupervised Text Style Transfercited by 12
  27. 2022Exploiting Transformer-based Multitask Learning for the Detection of Media Bias in News Articlescited by 11
  28. 2023The Media Bias Taxonomy: A Systematic Literature Review on the Forms and Automated Detection of Media Biascited by 11
  29. 2024SoMeLVLM: A Large Vision Language Model for Social Media Processingcited by 11
  30. 2021ENTRUST: Argument Reframing with Language Models and Entailmentcited by 10
  31. 2023To Revise or Not to Revise: Learning to Detect Improvable Claims for Argumentative Writing Supportcited by 8
  32. 2024Deciphering Political Entity Sentiment in News with Large Language Models: Zero-Shot and Few-Shot Strategiescited by 8
  33. 2024Seeing Like an AI: How LLMs Apply (and Misapply) Wikipedia Neutrality Normscited by 8
  34. 2021Mitigating Media Bias through Neutral Article Generationcited by 7
  35. 2024SOCIALITE-LLAMA: An Instruction-Tuned Model for Social Scientific Taskscited by 7
  36. 2024LLM-based Rewriting of Inappropriate Argumentation using Reinforcement Learning from Machine Feedbackcited by 7
  37. 2024Evaluating the Smooth Control of Attribute Intensity in Text Generation with LLMscited by 7
  38. 2021RECAST: Enabling User Recourse and Interpretability of Toxicity Detection Models with Interactive Visualizationcited by 6

50 most cited of 105 All 105 (JSON)