mlsanityData to 14 Oct 2026

Ex Machina: Personal Attacks Seen at Scale

arXiv 1610.08914 ↗First posted 27 Oct 2016cs.CLcited by 210

Not on the map. Why?

Abstract (arXiv)

The damage personal attacks cause to online discourse motivates many platforms to try to curb the phenomenon. However, understanding the prevalence and impact of personal attacks in online platforms at scale remains surprisingly difficult. The contribution of this paper is to develop and illustrate a method that combines crowdsourcing and machine learning to analyze personal attacks at scale. We show an evaluation method for a classifier in terms of the aggregated number of crowd-workers it can approximate. We apply our methodology to English Wikipedia, generating a corpus of over 100k high quality human-labeled comments and 63M machine-labeled ones from a classifier that is as good as the aggregate of 3 crowd-workers, as measured by the area under the ROC curve and Spearman correlation. Using this corpus of machine-labeled scores, our methodology allows us to explore some of the open questions about the nature of online personal attacks. This reveals that the majority of personal attacks on Wikipedia are not the result of a few malicious users, nor primarily the consequence of allowing anonymous contributions from unregistered users.

Key facts. Cited by 210 papers in the map (data through 14 Oct 2026). First posted on arXiv 27 Oct 2016.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2021: 49; 2022: 47; 2023: 40; 2024: 35; 2025: 21; 2026: 18.

210 papers.

Sort:

By area:

  1. 2023RWKV: Reinventing RNNs for the Transformer Eracited by 827
  2. 2021Annotators with Attitudes: How Annotator Beliefs And Identities Bias Toxic Language Detectioncited by 264
  3. 2023In-context Vectors: Making In Context Learning More Effective and Controllable Through Latent Space Steeringcited by 245
  4. 2023Large Language Model Alignment: A Surveycited by 243
  5. 2023ChatGPT: Jack of all trades, master of nonecited by 204
  6. 2022BlenderBot 3: a deployed conversational agent that continually learns to responsibly engagecited by 198
  7. 2022A New Generation of Perspective API: Efficient Multilingual Character-level Transformerscited by 185
  8. 2021Challenges in Detoxifying Language Modelscited by 169
  9. 2021ExT5: Towards Extreme Multi-Task Scaling for Transfer Learningcited by 161
  10. 2021Charformer: Fast Character Transformers via Gradient-based Subword Tokenizationcited by 160
  11. 2024Attacks, Defenses and Evaluations for LLM Conversation Safety: A Surveycited by 130
  12. 2021Scale Efficiently: Insights from Pre-training and Fine-tuning Transformerscited by 106
38 more
  1. 2022Plex: Towards Reliability using Pretrained Large Model Extensionscited by 102
  2. 2021Leveraging Sparse Linear Layers for Debuggable Deep Networkscited by 82
  3. 2021Uncertainty Baselines: Benchmarks for Uncertainty & Robustness in Deep Learningcited by 82
  4. 2021Anticipating Safety Issues in E2E Conversational AI: Framework and Toolingcited by 78
  5. 2022COLD: A Benchmark for Chinese Offensive Language Detectioncited by 76
  6. 2021On the Safety of Conversational Models: Taxonomy, Dataset, and Benchmarkcited by 73
  7. 2022Is Your Toxicity My Toxicity? Exploring the Impact of Rater Identity on Toxicity Annotationcited by 69
  8. 2022Human-AI Collaboration via Conditional Delegation: A Case Study of Content Moderationcited by 67
  9. 2021Towards generalisable hate speech detection: a review on obstacles and solutionscited by 66
  10. 2021"Will You Find These Shortcuts?" A Protocol for Evaluating the Faithfulness of Input Salience Methods for Text Classificationcited by 66
  11. 2022Handling Bias in Toxic Speech Detection: A Surveycited by 61
  12. 2022Detecting and Understanding Harmful Memes: A Surveycited by 58
  13. 2022Characteristics of Harmful Text: Towards Rigorous Benchmarking of Language Modelscited by 49
  14. 2023"HOT" ChatGPT: The promise of ChatGPT in detecting and discriminating hateful, offensive, and toxic comments on social mediacited by 45
  15. 2023Using In-Context Learning to Improve Dialogue Safetycited by 42
  16. 2022Quantifying Social Biases Using Templates is Unreliablecited by 39
  17. 2022The CRINGE Loss: Learning what language not to modelcited by 39
  18. 2022DIRECTOR: Generator-Classifiers For Supervised Language Modelingcited by 38
  19. 2022Toxicity Detection with Generative Prompt-based Inferencecited by 35
  20. 2021Detecting Harmful Content On Online Platforms: What Platforms Need Vs. Where Research Efforts Gocited by 33
  21. 2021Civil Rephrases Of Toxic Texts With Self-Supervised Transformerscited by 32
  22. 2022A Meta-Analysis of the Utility of Explainable Artificial Intelligence in Human-AI Decision-Makingcited by 29
  23. 2023Theoretical and Practical Perspectives on what Influence Functions Docited by 29
  24. 2021Are Pre-trained Convolutions Better than Pre-trained Transformers?cited by 27
  25. 2021A systematic review of Hate Speech automatic detection using Natural Language Processingcited by 27
  26. 2021Ruddit: Norms of Offensiveness for English Reddit Commentscited by 27
  27. 2023Disentangling Perceptions of Offensiveness: Cultural and Moral Correlatescited by 26
  28. 2024"They are uncultured": Unveiling Covert Harms and Social Threats in LLM Generated Conversationscited by 25
  29. 2021Conversations Gone Alright: Quantifying and Predicting Prosocial Outcomes in Online Conversationscited by 20
  30. 2024PolygloToxicityPrompts: Multilingual Evaluation of Neural Toxic Degeneration in Large Language Modelscited by 20
  31. 2023Make Every Example Count: On the Stability and Utility of Self-Influence for Learning from Noisy NLP Datasetscited by 19
  32. 2021I Beg to Differ: A study of constructive disagreement in online conversationscited by 18
  33. 2021Rethinking Search: Making Domain Experts out of Dilettantescited by 18
  34. 2021Measuring and Improving Model-Moderator Collaboration using Uncertainty Estimationcited by 18
  35. 2022Necessity and Sufficiency for Explaining Text Classifiers: A Case Study in Hate Speech Detectioncited by 18
  36. 2023GRASP: A Disagreement Analysis Framework to Assess Group Associations in Perspectivescited by 16
  37. 2022Proactive Moderation of Online Discussions: Existing Practices and the Potential for Algorithmic Supportcited by 15
  38. 2023Understanding Multi-Turn Toxic Behaviors in Open-Domain Chatbotscited by 15

50 most cited of 210 All 210 (JSON)