mlsanityData to 7 Oct 2026

Token-Specific Watermarking with Enhanced Detectability and Semantic Coherence for Large Language Models

arXiv 2402.18059 ↗First posted 28 Feb 2024cs.LG cs.CL cs.CRcited by 30

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

Abstract (arXiv)

Large language models generate high-quality responses with potential misinformation, underscoring the need for regulation by distinguishing AI-generated and human-written texts. Watermarking is pivotal in this context, which involves embedding hidden markers in texts during the LLM inference phase, which is imperceptible to humans. Achieving both the detectability of inserted watermarks and the semantic quality of generated texts is challenging. While current watermarking algorithms have made promising progress in this direction, there remains significant scope for improvement. To address these challenges, we introduce a novel multi-objective optimization (MOO) approach for watermarking that utilizes lightweight networks to generate token-specific watermarking logits and splitting ratios. By leveraging MOO to optimize for both detection and semantic objective functions, our method simultaneously achieves detectability and semantic integrity. Experimental results show that our method outperforms current watermarking techniques in enhancing the detectability of texts generated by LLMs while maintaining their semantic coherence. Our code is available at https://github.com/mignonjia/TS_watermark.

Key facts. Cited by 30 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 38 of its 62 references that are in the map. First posted on arXiv 28 Feb 2024.

Builds on

38 of 62 references.

Closest ancestors · 14

Same direction, newest first

  1. 2023Mark My Words: Analyzing and Evaluating Language Model Watermarks1 Dec
  2. 2023A Robust Semantics-based Watermark for Large Language Model against Paraphrasing15 Nov
  3. 2023REMARK-LLM: A Robust and Efficient Watermarking Framework for Generative Large Language Models18 Oct
  4. 2023A Semantic Invariant Robust Watermark for Large Language Models10 Oct
  5. 2023Robust Distortion-free Watermarks for Language Models28 Jul
  6. 2023Three Bricks to Consolidate Watermarks for Large Language Models26 Jul
  7. 2023On the Reliability of Watermarks for Large Language Models7 Jun
  8. 2023Who Wrote this Code? Watermarking for Code Generation24 May
  9. 2023Watermarking Text Generated by Black-Box Language Models14 May
  10. 2023Robust Multi-bit Natural Language Watermarking through Invariant Features3 May
  11. 2023Paraphrasing evades detectors of AI-generated text, but retrieval is an effective defense23 Mar
  12. 2023A Watermark for Large Language Models24 Jan
2 more
  1. 2022CATER: Intellectual Property Protection on Text Generation APIs via Conditional Watermarks19 Sep
  2. 2021Protecting Intellectual Property of Language Generation APIs with Lexical Watermark5 Dec

Other references · 24

  1. 2017Attention Is All You Needcited by 59,932
  2. 2020Language Models are Few-Shot Learnerscited by 32,199
  3. 2019Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformercited by 14,491
  4. 2019RoBERTa: A Robustly Optimized BERT Pretraining Approachcited by 13,266
  5. 2023Llama 2: Open Foundation and Fine-Tuned Chat Modelscited by 12,865
  6. 2015Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classificationcited by 3,590
  7. 2022OPT: Open Pre-trained Transformer Language Modelscited by 3,464
  8. 2019HuggingFace's Transformers: State-of-the-art Natural Language Processingcited by 2,201
  9. 2021SimCSE: Simple Contrastive Learning of Sentence Embeddingscited by 2,168
  10. 2017GradNorm: Gradient Normalization for Adaptive Loss Balancing in Deep Multitask Networkscited by 816
  11. 2018Multi-Task Learning as Multi-Objective Optimizationcited by 809
  12. 2019Release Strategies and the Social Impacts of Language Modelscited by 505
12 more
  1. 2019GLTR: Statistical Detection and Visualization of Generated Textcited by 490
  2. 2023How Close is ChatGPT to Human Experts? Comparison Corpus, Evaluation, and Detectioncited by 472
  3. 2019Automatic Detection of Generated Text is Easiest when Humans are Fooledcited by 260
  4. 2023GPT detectors are biased against non-native English writerscited by 175
  5. 2023Factuality Challenges in the Era of Large Language Modelscited by 125
  6. 2020Adversarial Watermarking Transformer: Towards Tracing Text Provenance with Data Hidingcited by 121
  7. 2023LLMDet: A Third Party Large Language Models Generated Text Detection Toolcited by 44
  8. 2020Neural Deepfake Detection with Factual Structure of Textcited by 40
  9. 2023Origin Tracing and Detecting of LLMscited by 39
  10. 2020Attacking Neural Text Detectorscited by 37
  11. 2019Multi-Gradient Descent for Multi-Objective Recommender Systemscited by 20
  12. 2023Composition-contrastive Learning for Sentence Embeddingscited by 14

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 12; 2025: 12; 2026: 6.

30 papers.

Sort:

By area:

  1. 2024WaterMax: breaking the LLM watermark detectability-robustness-quality trade-offcited by 48
  2. 2024SoK: Watermarking for AI-Generated Contentcited by 43
  3. 2024Watermarking Techniques for Large Language Models: A Surveycited by 20
  4. 2025AdaDetectGPT: Adaptive Detection of LLM-Generated Text with Statistical Guaranteescited by 20
  5. 2024Theoretically Grounded Framework for LLM Watermarking: A Distribution-Adaptive Approachcited by 17
  6. 2024A Watermark for Order-Agnostic Language Modelscited by 14
  7. 2024De-mark: Watermark Removal in Large Language Modelscited by 11
  8. 2024Watermarking Language Models with Error Correcting Codescited by 10
  9. 2024WaterSeeker: Pioneering Efficient Detection of Watermarked Segments in Large Documentscited by 9
  10. 2025Unified attacks to large language model watermarks: spoofing and scrubbing in unauthorized knowledge distillationcited by 9
  11. 2025In-Context Watermarks for Large Language Modelscited by 9
  12. 2024Topic-Based Watermarks for Large Language Modelscited by 7
18 more
  1. 2025Position: LLM Watermarking Should Align Stakeholders' Incentives for Practical Adoptioncited by 7
  2. 2024Survey on Plagiarism Detection in Large Language Models: The Impact of ChatGPT and Gemini on Academic Integritycited by 5
  3. 2025From Trade-off to Synergy: A Versatile Symbiotic Watermarking Framework for Large Language Modelscited by 5
  4. 2025Invisible Entropy: Towards Safe and Efficient Low-Entropy LLM Watermarkingcited by 5
  5. 2025SAEMark: Steering Personalized Multilingual LLM Watermarks with Sparse Autoencoderscited by 5
  6. 2025Factuality Beyond Coherence: Evaluating LLM Watermarking Methods for Medical Textscited by 4
  7. 2026A Visual Semantic Adaptive Watermark grounded by Prefix-Tuning for Large Vision-Language Modelcited by 3
  8. 2024Watermarking Large Language Models and the Generated Content: Opportunities and Challengescited by 2
  9. 2025SoK: Are Watermarks in LLMs Ready for Deployment?cited by 2
  10. 2025Learning to Watermark: A Selective Watermarking Framework for Large Language Models via Multi-Objective Optimizationcited by 1
  11. 2026SearchLLM: Detecting LLM Paraphrased Text by Measuring the Similarity with Regeneration of the Candidate Source via Search Enginecited by 1
  12. 2024Less is More: Sparse Watermarking in LLMs with Enhanced Text Qualitycited by 0
  13. 2025Robust and Secure Code Watermarking for Large Language Models via ML/Crypto Codesigncited by 0
  14. 2025Adaptive Testing for Segmenting Watermarked Texts From Language Modelscited by 0
  15. 2026On Google's SynthID-Text LLM Watermarking System: Theoretical Analysis and Empirical Validationcited by 0
  16. 2026Watermarking Game-Playing Agents in Perfect-Information Extensive-Form Gamescited by 0
  17. 2026READER: Reasoning-Enhanced AI-Generated Text Detectioncited by 0
  18. 2026A Locally Tokenized Generative Model for Robust Time-Series Watermarkingcited by 0