mlsanityData to 7 Oct 2026

Watermarking generated content and models

717 papers6 this weekHigh confidence in this grouping: the papers form one clear topicHigh confidence in this grouping: the papers form one clear topic · About

Keywords: watermarking, watermark, watermarks, text, detection

Key facts. Watermarking generated content and models: 717 papers on arXiv since 2021 (data through 7 Oct 2026). 282 of its 717 mapped papers (39%) cite “A Watermark for Large Language Models” (2023, arXiv 2301.10226). 227 of its 717 mapped papers (32%) cite “Robust Distortion-free Watermarks for Language Models” (2023, arXiv 2307.15593). 21 new papers 24 Sep–7 Oct.

What’s new

24 Sep – 7 Oct 2026

21 new papers, 24 Sep – 7 Oct; 6 this week.

  1. This week · 6
  2. Diffusion-Generated Image Watermarking: A Two-Axis Taxonomy and Three Protocol-Bounded Case Studies
  3. Latent Watermarks under Generative Editing: A Benchmark and Analysis of Detection Survival
  4. Semantic Behavioral Watermarking: Paraphrase-Robust and Forgery-Resistant Provenance for LLM Agents
  5. MARCO: The Radioactive Watermark for Protein Generative Models
  6. Rethinking Visual Provenance: Detection and Watermarking Across Direct Visual Generation and LLM-Driven Code Rendering
  7. SimpleMark: Fast Multi-Bit Text Watermarking under f -Divergence Constraints
  8. Earlier · 15
  9. The Poisoned Conversation: Privacy-Leaking Watermarks in Unified Multimodal Models
  10. Invisible Ink, Visible Lies: How Production Watermarking Causes LLMs to Hallucinate
  11. Language Model Fingerprinting Requires Rethinking Watermark Teachers
  12. RMCW: A Deletion-Robust Watermark Based on Reed--Muller Codes for Language Models
All 21
  1. Exploring Weaknesses of Generative Image Watermarks against Latent Frequency Masking
  2. Persistent Watermarking of Text-to-Image Models
  3. Anchor-ECC: Local Integrity Checking for Watermarked LLM Outputs via Error-Correcting Codes
  4. AutoMark: Enabling Autoresearch to Discover Better LLM Watermarks
  5. Beyond Semantic Narrowing: Robust and Efficient LLM Watermarking with Hamming Neighborhoods
  6. TTMark: Pairwise Distortion-Free Watermarking Beyond Single-Token Entropy
  7. TANGO: Watermarking Masked Diffusion Language Models in Token Pairs
  8. GenoTrace: Inheritable Watermarks for Genome Foundation Model Distillation
  9. Tokens Change, Structure Endures: Spectral Watermarking for Generated Speech
  10. CertMark: Distortion-Free Multi-Bit Watermarking with Certified Decoding
  11. FeatMark: Feature-level Watermark Protection against Mimicry Attacks with Diffusion Models

Foundations

Most cited by papers here

  1. A Watermark for Large Language Models2023cited by 282 here
  2. Robust Distortion-free Watermarks for Language Models2023cited by 227 here
  3. Provable Robust Watermarking for AI-Generated Text2023cited by 205 here
  4. Undetectable Watermarks for Language Models2023cited by 148 here
  5. The Stable Signature: Rooting Watermarks in Latent Diffusion Models2023cited by 145 here

Lineage

The papers this direction builds on, oldest first.

Each paper on the line cites the one before.

  1. 2023A Watermark for Large Language Modelscited by 282 here
  2. 2023Can AI-Generated Text be Reliably Detected?cited by 90 here
  3. 2023Paraphrasing evades detectors of AI-generated text, but retrieval is an effective defensecited by 137 here
  4. 2023Undetectable Watermarks for Language Modelscited by 148 here
  5. 2023Robust Distortion-free Watermarks for Language Modelscited by 227 here
  6. 2023A Semantic Invariant Robust Watermark for Large Language Modelscited by 90 here
  7. 2023On the Learnability of Watermarks for Language Modelscited by 56 here
  8. 2023A Survey of Text Watermarking in the Era of Large Language Modelscited by 72 here
  9. 2024An Entropy-based Text Watermarking Detection Methodcited by 57 here
  10. 2024MarkLLM: An Open-Source Toolkit for LLM Watermarkingcited by 75 here
14 more it builds on, not in this chain
  1. 2023A Recipe for Watermarking Diffusion Modelscited by 65 here
  2. 2023The Stable Signature: Rooting Watermarks in Latent Diffusion Modelscited by 145 here
  3. 2023Who Wrote this Code? Watermarking for Code Generationcited by 108 here
  4. 2023Tree-Ring Watermarks: Fingerprints for Diffusion Images that are Invisible and Robustcited by 85 here
  5. 2023Invisible Image Watermarks Are Provably Removable Using Generative AIcited by 73 here
  6. 2023On the Reliability of Watermarks for Large Language Modelscited by 119 here
  7. 2023Provable Robust Watermarking for AI-Generated Textcited by 205 here
  8. 2023Three Bricks to Consolidate Watermarks for Large Language Modelscited by 65 here
  9. 2023Unbiased Watermark for Large Language Modelscited by 87 here
  10. 2023SemStamp: A Semantic Watermark with Paraphrastic Robustness for Text Generationcited by 99 here
  11. 2023Publicly-Detectable Watermarking for Language Modelscited by 61 here
  12. 2023A Robust Semantics-based Watermark for Large Language Model against Paraphrasingcited by 67 here
  13. 2024Watermark Stealing in Large Language Modelscited by 66 here
  14. 2024Gaussian Shading: Provable Performance-Lossless Image Watermarking for Diffusion Modelscited by 90 here

The 24 most cited of 677.

Browse

30 of 717, ranked 6–35

  1. 6Paraphrasing evades detectors of AI-generated text, but retrieval is an effective defense2023137
  2. 7On the Reliability of Watermarks for Large Language Models2023119
  3. 8Who Wrote this Code? Watermarking for Code Generation2023108
  4. 9SemStamp: A Semantic Watermark with Paraphrastic Robustness for Text Generation202399
  5. 10Gaussian Shading: Provable Performance-Lossless Image Watermarking for Diffusion Models202490
  6. 11Can AI-Generated Text be Reliably Detected?202390
  7. 12A Semantic Invariant Robust Watermark for Large Language Models202390
  8. 13Unbiased Watermark for Large Language Models202387
  9. 14Tree-Ring Watermarks: Fingerprints for Diffusion Images that are Invisible and Robust202385
  10. 15MarkLLM: An Open-Source Toolkit for LLM Watermarking202475
All 30
  1. 16Invisible Image Watermarks Are Provably Removable Using Generative AI202373
  2. 17A Survey of Text Watermarking in the Era of Large Language Models202372
  3. 18A Robust Semantics-based Watermark for Large Language Model against Paraphrasing202367
  4. 19Watermark Stealing in Large Language Models202466
  5. 20A Recipe for Watermarking Diffusion Models202365
  6. 21Three Bricks to Consolidate Watermarks for Large Language Models202365
  7. 22Publicly-Detectable Watermarking for Language Models202361
  8. 23An Entropy-based Text Watermarking Detection Method202457
  9. 24On the Learnability of Watermarks for Language Models202356
  10. 25Robust Multi-bit Natural Language Watermarking through Invariant Features202355
  11. 26Mark My Words: Analyzing and Evaluating Language Model Watermarks202355
  12. 27REMARK-LLM: A Robust and Efficient Watermarking Framework for Generative Large Language Models202354
  13. 28Adaptive Text Watermark for Large Language Models202453
  14. 29Protecting Language Generation Models via Invisible Watermarking202352
  15. 30An Undetectable Watermark for Generative Image Models202450
  16. 31An Unforgeable Publicly Verifiable Watermark for Large Language Models202350
  17. 32Watermarks in the Sand: Impossibility of Strong Watermarking for Generative Models202350
  18. 33Protecting Intellectual Property of Language Generation APIs with Lexical Watermark202148
  19. 34Advancing Beyond Identification: Multi-bit Watermark for Large Language Models202347
  20. 35Watermarking Text Generated by Black-Box Language Models202346