mlsanityData to 7 Oct 2026

Three Bricks to Consolidate Watermarks for Large Language Models

arXiv 2308.00113 ↗First posted 26 Jul 2023cs.CL cs.AIcited by 80

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

Abstract (arXiv)

The task of discerning between generated and natural texts is increasingly challenging. In this context, watermarking emerges as a promising technique for ascribing generated text to a specific model. It alters the sampling generation process so as to leave an invisible trace in the generated output, facilitating later detection. This research consolidates watermarks for large language models based on three theoretical and empirical considerations. First, we introduce new statistical tests that offer robust theoretical guarantees which remain valid even at low false-positive rates (less than 10-6). Second, we compare the effectiveness of watermarks using classical benchmarks in the field of natural language processing, gaining insights into their real-world applicability. Third, we develop advanced detection schemes for scenarios where access to the LLM is available, as well as multi-bit watermarking.

Key facts. Cited by 80 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 22 of its 36 references that are in the map. First posted on arXiv 26 Jul 2023.

Builds on

22 of 36 references.

Closest ancestors · 7

Same direction, newest first

  1. 2023Provable Robust Watermarking for AI-Generated Text30 Jun
  2. 2023On the Reliability of Watermarks for Large Language Models7 Jun
  3. 2023Tree-Ring Watermarks: Fingerprints for Diffusion Images that are Invisible and Robust31 May
  4. 2023Undetectable Watermarks for Language Models25 May
  5. 2023The Stable Signature: Rooting Watermarks in Latent Diffusion Models27 Mar
  6. 2023Protecting Language Generation Models via Invisible Watermarking6 Feb
  7. 2023A Watermark for Large Language Models24 Jan

Other references · 15

  1. 2023LLaMA: Open and Efficient Foundation Language Modelscited by 13,958
  2. 2021Training Verifiers to Solve Math Word Problemscited by 9,743
  3. 2021Evaluating Large Language Models Trained on Codecited by 8,053
  4. 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,203
  5. 2021Program Synthesis with Large Language Modelscited by 3,509
  6. 2023QLoRA: Efficient Finetuning of Quantized LLMscited by 3,355
  7. 2017TriviaQA: A Large Scale Distantly Supervised Challenge Dataset for Reading Comprehensioncited by 2,895
  8. 2019The Curious Case of Neural Text Degenerationcited by 2,881
  9. 2018Hierarchical Neural Story Generationcited by 1,182
  10. 2023DetectGPT: Zero-Shot Machine-Generated Text Detection using Probability Curvaturecited by 705
  11. 2019Defending Against Neural Fake Newscited by 553
  12. 2019Automatic Detection of Generated Text is Easiest when Humans are Fooledcited by 260
3 more
  1. 2020Responsible Disclosure of Generative Models Using Scalable Fingerprintingcited by 90
  2. 2022Machine Generated Text: A Comprehensive Survey of Threat Models and Detection Methodscited by 75
  3. 2023The Age of Synthetic Realities: Challenges and Opportunitiescited by 15

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 7; 2024: 21; 2025: 25; 2026: 27.

80 papers.

Sort:

By area:

  1. 2024Monitoring AI-Modified Content at Scale: A Case Study on the Impact of ChatGPT on AI Conference Peer Reviewscited by 160
  2. 2024Mapping the Increasing Use of LLMs in Scientific Paperscited by 112
  3. 2024Detecting Multimedia Generated by Large AI Models: A Surveycited by 73
  4. 2024Proactive Detection of Voice Cloning with Localized Watermarkingcited by 67
  5. 2023Preventing Language Models From Hiding Their Reasoningcited by 62
  6. 2023Mark My Words: Analyzing and Evaluating Language Model Watermarkscited by 62
  7. 2024Watermarking Makes Language Models Radioactivecited by 52
  8. 2023Advancing Beyond Identification: Multi-bit Watermark for Large Language Modelscited by 51
  9. 2024WaterMax: breaking the LLM watermark detectability-robustness-quality trade-offcited by 48
  10. 2023Towards Codable Watermarking for Injecting Multi-bits Information to LLMscited by 43
  11. 2024A Watermark-Conditioned Diffusion Model for IP Protectioncited by 34
  12. 2024Token-Specific Watermarking with Enhanced Detectability and Semantic Coherence for Large Language Modelscited by 30
38 more
  1. 2023Towards Possibilities & Impossibilities of AI-generated Text Detection: A Surveycited by 28
  2. 2023From Text to Source: Results in Detecting Large Language Model-Generated Contentcited by 25
  3. 2023Towards Optimal Statistical Watermarkingcited by 25
  4. 2024Learning to Watermark LLM-generated Text via Reinforcement Learningcited by 19
  5. 2024Debiasing Watermarks for Large Language Models via Maximal Couplingcited by 18
  6. 2025Detecting LLM-Generated Peer Reviewscited by 18
  7. 2025BiMark: Unbiased Multilayer Watermarking for Large Language Modelscited by 18
  8. 2024Permute-and-Flip: An optimally stable and watermarkable decoder for LLMscited by 17
  9. 2024Theoretically Grounded Framework for LLM Watermarking: A Distribution-Adaptive Approachcited by 17
  10. 2025Improving Your Model Ranking on Chatbot Arena by Vote Riggingcited by 16
  11. 2025Watermarking Autoregressive Image Generationcited by 15
  12. 2024A Watermark for Black-Box Language Modelscited by 14
  13. 2024Excuse me, sir? Your language model is leaking (information)cited by 11
  14. 2025StealthInk: A Multi-bit and Stealthy Watermark for Large Language Modelscited by 11
  15. 2024Multi-Bit Distortion-Free Watermarking for Large Language Modelscited by 10
  16. 2025Optimized Couplings for Watermarking Large Language Modelscited by 9
  17. 2025In-Context Watermarks for Large Language Modelscited by 9
  18. 2025Watermarking Diffusion Language Modelscited by 9
  19. 2024Inevitable Trade-off between Watermark Strength and Speculative Sampling Efficiency for Language Modelscited by 7
  20. 2025LLM Fingerprinting via Semantically Conditioned Watermarkscited by 7
  21. 2025Position: LLM Watermarking Should Align Stakeholders' Incentives for Practical Adoptioncited by 7
  22. 2024WaterPool: A Watermark Mitigating Trade-offs among Imperceptibility, Efficacy and Robustnesscited by 6
  23. 2024Efficiently Identifying Watermarked Segments in Mixed-Source Textscited by 6
  24. 2025Detecting Benchmark Contamination Through Watermarkingcited by 6
  25. 2026Antidistillation Fingerprintingcited by 6
  26. 2025On the Empirical Power of Goodness-of-Fit Tests in Watermark Detectioncited by 5
  27. 2026Tool Use Enables Undetectable Steganography in Multi-Agent LLM Systemscited by 5
  28. 2024Discovering Spoofing Attempts on Language Model Watermarkscited by 4
  29. 2024Segmenting Watermarked Texts From Language Modelscited by 4
  30. 2025Can you Finetune your Binoculars? Embedding Text Watermarks into the Weights of Large Language Modelscited by 4
  31. 2025BiMarker: Enhancing Text Watermark Detection for Large Language Models with Bipolar Watermarkscited by 3
  32. 2025Multi-use LLM Watermarking and the False Detection Problemcited by 3
  33. 2025Majority Bit-Aware Watermarking For Large Language Modelscited by 3
  34. 2025Detecting Post-generation Edits to Watermarked LLM Outputs via Combinatorial Watermarkingcited by 3
  35. 2026A Unified Framework for LLM Watermarkscited by 3
  36. 2025A Statistical Hypothesis Testing Framework for Data Misappropriation Detection in Large Language Modelscited by 2
  37. 2025How Good is Post-Hoc Watermarking With Language Model Rephrasing?cited by 2
  38. 2026MC2Mark: Distortion-Free Multi-Bit Watermarking for Long Messagescited by 2

50 most cited of 80 All 80 (JSON)