mlsanityData to 7 Oct 2026

Provable Robust Watermarking for AI-Generated Text

arXiv 2306.17439 ↗First posted 30 Jun 2023cs.CL cs.LGcited by 259

On the map: LLM safety and security → Watermarking generated content and models

Abstract (arXiv)

We study the problem of watermarking large language models (LLMs) generated text -- one of the most promising approaches for addressing the safety challenges of LLM usage. In this paper, we propose a rigorous theoretical framework to quantify the effectiveness and robustness of LLM watermarks. We propose a robust and high-quality watermark method, Unigram-Watermark, by extending an existing approach with a simplified fixed grouping strategy. We prove that our watermark method enjoys guaranteed generation quality, correctness in watermark detection, and is robust against text editing and paraphrasing. Experiments on three varying LLMs and two datasets verify that our Unigram-Watermark achieves superior detection accuracy and comparable generation quality in perplexity, thus promoting the responsible use of LLMs. Code is available at https://github.com/XuandongZhao/Unigram-Watermark.

Key facts. Cited by 259 papers in the map (data through 7 Oct 2026). Placed in Watermarking generated content and models, one of 12 directions in LLM safety and security. Builds on 31 of its 59 references that are in the map. First posted on arXiv 30 Jun 2023.

Builds on

31 of 59 references.

Closest ancestors · 6

Same direction, newest first

  1. 2023Undetectable Watermarks for Language Models25 May
  2. 2023Paraphrasing evades detectors of AI-generated text, but retrieval is an effective defense23 Mar
  3. 2023Can AI-Generated Text be Reliably Detected?17 Mar
  4. 2023Protecting Language Generation Models via Invisible Watermarking6 Feb
  5. 2023A Watermark for Large Language Models24 Jan
  6. 2021Tracing Text Provenance via Context-Aware Lexical Substitution15 Dec

Other references · 25

  1. 2017Attention Is All You Needcited by 59,932
  2. 2020Language Models are Few-Shot Learnerscited by 32,199
  3. 2023GPT-4 Technical Reportcited by 18,132
  4. 2022Training language models to follow instructions with human feedbackcited by 15,712
  5. 2019Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformercited by 14,491
  6. 2023LLaMA: Open and Efficient Foundation Language Modelscited by 13,958
  7. 2019RoBERTa: A Robustly Optimized BERT Pretraining Approachcited by 13,266
  8. 2022Hierarchical Text-Conditional Image Generation with CLIP Latentscited by 6,539
  9. 2022Photorealistic Text-to-Image Diffusion Models with Deep Language Understandingcited by 6,322
  10. 2019BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehensioncited by 5,874
  11. 2022Robust Speech Recognition via Large-Scale Weak Supervisioncited by 3,972
  12. 2022OPT: Open Pre-trained Transformer Language Modelscited by 3,464
13 more
  1. 2016Pointer Sentinel Mixture Modelscited by 2,992
  2. 2019The Curious Case of Neural Text Degenerationcited by 2,881
  3. 2019HuggingFace's Transformers: State-of-the-art Natural Language Processingcited by 2,201
  4. 2015Teaching Machines to Read and Comprehendcited by 1,039
  5. 2021Ethical and social risks of harm from Language Modelscited by 957
  6. 2023DetectGPT: Zero-Shot Machine-Generated Text Detection using Probability Curvaturecited by 705
  7. 2019Defending Against Neural Fake Newscited by 553
  8. 2019GLTR: Statistical Detection and Visualization of Generated Textcited by 490
  9. 2023Poisoning Web-Scale Training Datasets is Practicalcited by 276
  10. 2023GPT detectors are biased against non-native English writerscited by 175
  11. 2023On the Possibilities of AI-Generated Text Detectioncited by 88
  12. 2020Attacking Neural Text Detectorscited by 37
  13. 2021Frustratingly Easy Edit-based Linguistic Steganography with a Masked Language Modelcited by 28

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 25; 2024: 81; 2025: 83; 2026: 70.

259 papers.

Sort:

By area:

  1. 2023Can LLM-Generated Misinformation Be Detected?cited by 173
  2. 2024Monitoring AI-Modified Content at Scale: A Case Study on the Impact of ChatGPT on AI Conference Peer Reviewscited by 160
  3. 2024Mapping the Increasing Use of LLMs in Scientific Paperscited by 112
  4. 2023A Survey of Text Watermarking in the Era of Large Language Modelscited by 102
  5. 2023A Semantic Invariant Robust Watermark for Large Language Modelscited by 97
  6. 2023Identifying and Mitigating the Security Risks of Generative AIcited by 90
  7. 2023Three Bricks to Consolidate Watermarks for Large Language Modelscited by 80
  8. 2024MarkLLM: An Open-Source Toolkit for LLM Watermarkingcited by 80
  9. 2024Watermark Stealing in Large Language Modelscited by 79
  10. 2023Watermarks in the Sand: Impossibility of Strong Watermarking for Generative Modelscited by 77
  11. 2024Detecting Multimedia Generated by Large AI Models: A Surveycited by 73
  12. 2023On the Learnability of Watermarks for Language Modelscited by 72
38 more
  1. 2023A Survey on Detection of LLMs-Generated Contentcited by 70
  2. 2023REMARK-LLM: A Robust and Efficient Watermarking Framework for Generative Large Language Modelscited by 64
  3. 2024Many Heads Are Better Than One: Improved Scientific Idea Generation by A LLM-Based Multi-Agent Systemcited by 63
  4. 2023Mark My Words: Analyzing and Evaluating Language Model Watermarkscited by 62
  5. 2024An Entropy-based Text Watermarking Detection Methodcited by 61
  6. 2024Adaptive Text Watermark for Large Language Modelscited by 60
  7. 2023An Unforgeable Publicly Verifiable Watermark for Large Language Modelscited by 54
  8. 2024WaterMax: breaking the LLM watermark detectability-robustness-quality trade-offcited by 48
  9. 2023A Resilient and Accessible Distribution-Preserving Watermark for Large Language Modelscited by 47
  10. 2024No Free Lunch in LLM Watermarking: Trade-offs in Watermarking Design Choicescited by 47
  11. 2024k-SemStamp: A Clustering-Based Semantic Watermark for Detection of Machine-Generated Textcited by 46
  12. 2023Towards Codable Watermarking for Injecting Multi-bits Information to LLMscited by 43
  13. 2024M4GT-Bench: Evaluation Benchmark for Black-Box Machine-Generated Text Detectioncited by 43
  14. 2024SoK: Watermarking for AI-Generated Contentcited by 43
  15. 2024Pseudorandom Error-Correcting Codescited by 40
  16. 2023Fighting Fire with Fire: Can ChatGPT Detect AI-generated Text?cited by 39
  17. 2024Copyright Protection in Generative AI: A Technical Perspectivecited by 35
  18. 2025Are You Getting What You Pay For? Auditing Model Substitution in LLM APIscited by 35
  19. 2024GumbelSoft: Diversified Language Model Watermarking via the GumbelMax-trickcited by 31
  20. 2024On the Challenges and Opportunities in Generative AIcited by 31
  21. 2023Towards Possibilities & Impossibilities of AI-generated Text Detection: A Surveycited by 28
  22. 2023WaterBench: Towards Holistic Evaluation of Watermarks for Large Language Modelscited by 27
  23. 2024Can Watermarks Survive Translation? On the Cross-lingual Consistency of Text Watermark for Large Language Modelscited by 27
  24. 2023Towards Optimal Statistical Watermarkingcited by 25
  25. 2024Stumbling Blocks: Stress Testing the Robustness of Machine-Generated Text Detectors Under Attackscited by 25
  26. 2023Grounding Foundation Models through Federated Transfer Learning: A General Frameworkcited by 24
  27. 2024Bypassing LLM Watermarks with Color-Aware Substitutionscited by 24
  28. 2024PostMark: A Robust Blackbox Watermark for Large Language Modelscited by 24
  29. 2023Embarrassingly Simple Text Watermarkscited by 23
  30. 2024Detecting AI-Generated Text: Factors Influencing Detectability with Current Methodscited by 23
  31. 2023WatME: Towards Lossless Watermarking Through Lexical Redundancycited by 22
  32. 2023Optimizing watermarks for large language modelscited by 22
  33. 2024Bileve: Securing Text Provenance in Large Language Models Against Spoofing with Bi-level Signaturecited by 21
  34. 2024Watermarking Techniques for Large Language Models: A Surveycited by 20
  35. 2025AdaDetectGPT: Adaptive Detection of LLM-Generated Text with Statistical Guaranteescited by 20
  36. 2026Large Language Model Reasoning Failurescited by 20
  37. 2024Learning to Watermark LLM-generated Text via Reinforcement Learningcited by 19
  38. 2024Revisiting the Robustness of Watermarking to Paraphrasing Attackscited by 19

50 most cited of 259 All 259 (JSON)