mlsanityData to 14 Oct 2026

An Unforgeable Publicly Verifiable Watermark for Large Language Models

arXiv 2307.16230 ↗First posted 30 Jul 2023cs.CLcited by 65

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

Abstract (arXiv)

Recently, text watermarking algorithms for large language models (LLMs) have been proposed to mitigate the potential harms of text generated by LLMs, including fake news and copyright issues. However, current watermark detection algorithms require the secret key used in the watermark generation process, making them susceptible to security breaches and counterfeiting during public detection. To address this limitation, we propose an unforgeable publicly verifiable watermark algorithm named UPV that uses two different neural networks for watermark generation and detection, instead of using the same key at both stages. Meanwhile, the token embedding parameters are shared between the generation and detection networks, which makes the detection network achieve a high accuracy very efficiently. Experiments demonstrate that our algorithm attains high detection accuracy and computational efficiency through neural networks. Subsequent analysis confirms the high complexity involved in forging the watermark from the detection network. Our code is available at hrefhttps://github.com/THU-BPM/unforgeablewatermarkhttps://github.com/THU-BPM/unforgeablewatermark. Additionally, our algorithm could also be accessed through MarkLLM citeppan2024markllm footnotehttps://github.com/THU-BPM/MarkLLM.

Key facts. Cited by 65 papers in the map (data through 14 Oct 2026). Placed in Watermarking generated content and models, one of 12 directions in LLM safety and security. Builds on 23 of its 33 references that are in the map. First posted on arXiv 30 Jul 2023.

Builds on

23 of 33 references.

Closest ancestors · 7

Same direction, newest first

  1. 2023Robust Distortion-free Watermarks for Language Models28 Jul
  2. 2023Provable Robust Watermarking for AI-Generated Text30 Jun
  3. 2023Undetectable Watermarks for Language Models25 May
  4. 2023Who Wrote this Code? Watermarking for Code Generation24 May
  5. 2023Robust Multi-bit Natural Language Watermarking through Invariant Features3 May
  6. 2023Can AI-Generated Text be Reliably Detected?17 Mar
  7. 2023A Watermark for Large Language Models24 Jan

Other references · 16

  1. 2019Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformercited by 14,599
  2. 2023LLaMA: Open and Efficient Foundation Language Modelscited by 14,132
  3. 2019RoBERTa: A Robustly Optimized BERT Pretraining Approachcited by 13,372
  4. 2019BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehensioncited by 5,893
  5. 2022OPT: Open Pre-trained Transformer Language Modelscited by 3,512
  6. 2022No Language Left Behind: Scaling Human-Centered Machine Translationcited by 1,312
  7. 2023DetectLLM: Leveraging Log Rank Information for Zero-Shot Detection of Machine-Generated Textcited by 210
  8. 2023RADAR: Robust AI-Text Detection via Adversarial Learningcited by 190
  9. 2020Adversarial Watermarking Transformer: Towards Tracing Text Provenance with Data Hidingcited by 130
  10. 2023On the Risk of Misinformation Pollution with Large Language Modelscited by 128
  11. 2023A comprehensive evaluation of ChatGPT's zero-shot Text-to-SQL capabilitycited by 119
  12. 2023On the Possibilities of AI-Generated Text Detectioncited by 88
4 more
  1. 2023A Pathway Towards Responsible AI Generated Contentcited by 38
  2. 2023Smaller Language Models are Better Black-box Machine-Generated Text Detectorscited by 35
  3. 2022Character-level White-Box Adversarial Attacks against Transformers via Attachable Subwords Substitutioncited by 28
  4. 2023G3Detector: General GPT-Generated Text Detectorcited by 17

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2023: 5; 2024: 18; 2025: 23; 2026: 19.

65 papers.

Sort:

By area:

  1. 2023A Survey of Text Watermarking in the Era of Large Language Modelscited by 112
  2. 2023A Survey on LLM-Generated Text Detection: Necessity, Methods, and Future Directionscited by 105
  3. 2024Detecting Multimedia Generated by Large AI Models: A Surveycited by 76
  4. 2023Mark My Words: Analyzing and Evaluating Language Model Watermarkscited by 71
  5. 2024Adaptive Text Watermark for Large Language Modelscited by 67
  6. 2024SoK: Watermarking for AI-Generated Contentcited by 51
  7. 2024Provably Robust Multi-bit Watermarking for AI-generated Textcited by 47
  8. 2023Towards Possibilities & Impossibilities of AI-generated Text Detection: A Surveycited by 28
  9. 2023Optimizing watermarks for large language modelscited by 25
  10. 2024Double-I Watermark: Protecting Model Copyright for LLM Fine-tuningcited by 21
  11. 2025Mark Your LLM: Detecting the Misuse of Open-Source Large Language Models via Watermarkingcited by 19
  12. 2025Revealing Weaknesses in Text Watermarking Through Self-Information Rewrite Attackscited by 18
38 more
  1. 2024A Watermark for Black-Box Language Modelscited by 16
  2. 2024A Watermark for Order-Agnostic Language Modelscited by 14
  3. 2024SoK: On the Role and Future of AIGC Watermarking in the Era of Gen-AIcited by 14
  4. 2024Can Watermarked LLMs be Identified by Users via Crafted Prompts?cited by 13
  5. 2024Distortion-free Watermarks are not Truly Distortion-free under Watermark Key Collisionscited by 10
  6. 2025In-Context Watermarks for Large Language Modelscited by 10
  7. 2025Align is not Enough: Multimodal Universal Jailbreak Attack against Multimodal Large Language Modelscited by 9
  8. 2025Enhancing LLM Watermark Resilience Against Both Scrubbing and Spoofing Attackscited by 9
  9. 2024PersonaMark: Personalized LLM watermarking for model protection and user attributioncited by 7
  10. 2025Position: LLM Watermarking Should Align Stakeholders' Incentives for Practical Adoptioncited by 7
  11. 2024Discovering Spoofing Attempts on Language Model Watermarkscited by 6
  12. 2024FoldMark: Protecting Protein Generative Models with Watermarkingcited by 6
  13. 2024Building Intelligence Identification System via Large Language Model Watermarking: A Survey and Beyondcited by 5
  14. 2025Paladin: Defending LLM-enabled Phishing Emails with a New Trigger-Tag Paradigmcited by 5
  15. 2025MarkDiffusion: An Open-Source Toolkit for Generative Watermarking of Latent Diffusion Modelscited by 5
  16. 2025Analyzing and Evaluating Unbiased Language Model Watermarkcited by 5
  17. 2025Large Language Model Sourcing: A Surveycited by 5
  18. 2024On Evaluating The Performance of Watermarked Machine-Generated Texts Under Adversarial Attackscited by 4
  19. 2025An End-to-End Model for Logits-Based Large Language Models Watermarkingcited by 4
  20. 2025A Crack in the Bark: Leveraging Public Knowledge to Remove Tree-Ring Watermarkscited by 4
  21. 2025Majority Bit-Aware Watermarking For Large Language Modelscited by 4
  22. 2026MC2Mark: Distortion-Free Multi-Bit Watermarking for Long Messagescited by 4
  23. 2024A Certified Robust Watermark For Large Language Modelscited by 3
  24. 2025Enhancing Watermarking Quality for LLMs via Contextual Generation States Awarenesscited by 3
  25. 2026Distilling the Thought, Watermarking the Answer: A Principle Semantic Guided Watermark for Large Reasoning Modelscited by 3
  26. 2026A Visual Semantic Adaptive Watermark grounded by Prefix-Tuning for Large Vision-Language Modelcited by 3
  27. 2026MirrorMark: Generalizable Mirrored Sampling for Multi-bit LLM Watermarkingcited by 3
  28. 2025SoK: Are Watermarks in LLMs Ready for Deployment?cited by 2
  29. 2025RLCracker: Evaluating the Worst-Case Vulnerability of LLM Watermarks with Adaptive RL Attackscited by 2
  30. 2025PVMark: Enabling Public Verifiability for LLM Watermarking Schemescited by 2
  31. 2026Refined Detection for Gumbel Watermarkingcited by 2
  32. 2026Watermarking Should Be Treated as a Monitoring Primitivecited by 2
  33. 2026SAMark: A Self-Anchored Text Watermarking with Paragraph-Level Paraphrase Robustnesscited by 2
  34. 2026Hidden in Plain Tokens: Simply Robust, Gradient-Free Watermark for Synthetic Audiocited by 2
  35. 2024MCGMark: An Encodable and Robust Online Watermark for Tracing LLM-Generated Malicious Codecited by 1
  36. 2025LLM Watermark Evasion via Bias Inversioncited by 1
  37. 2026VOW: Verifiable and Oblivious Watermark Detection for Large Language Modelscited by 1
  38. 2026LLM Watermarking as Big Data Provenance: A Deployment-Oriented Systematizationcited by 1

50 most cited of 65 All 65 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2026Unforgeable Watermarks for Language Models via Robust Signatures
  2. 2024Provably Robust Watermarks for Open-Source Language Models
  3. 2024Let Watermarks Speak: A Robust and Unforgeable Watermark for Language Models
  4. 2023Advancing Beyond Identification: Multi-bit Watermark for Large Language Models
  5. 2025PRO: Enabling Precise and Robust Text Watermark for Open-Source LLMs
  6. 2023Necessary and Sufficient Watermark for Large Language Models
  7. 2023On the Learnability of Watermarks for Language Models
  8. 2025Security and Detectability Analysis of Unicode Text Watermarking Methods against Large Language Models