Research map / Large language models / Language, safety and interpretability
WaterMod: Modular Token-Rank Partitioning for Probability-Balanced LLM Watermarking
Shinwoo Park, Hyejin Park, Hyeseon Ahn, Yo-Sub Han
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use rank-modular partition (WaterMod) instead of SynthID-Text | zero-bit watermark · GSM8K | accuracy: 47.61 -> 53.83; 13.06% | Improved |
| use rank-modular partition (WaterMod) instead of KGW | zero-bit watermark · MBPP+ | AUROC: 72.43 -> 82.66; 14.12% | Improved |
| use rank-modular partition (WaterMod) instead of KGW | zero-bit watermark · MBPP+ | pass@1: 29.90 -> 36.80; 23.07% | Improved |
| use rank-modular partition (WaterMod) instead of LSH | zero-bit watermark · MBPP+ | AUROC: 30.72 -> 82.66; 169.07% | Improved |
| use rank-modular partition (WaterMod) instead of LSH | zero-bit watermark · C4 | perplexity: 26.19 -> 12.58 | Improved |
| use rank-modular partition (WaterMod) instead of LSH | zero-bit watermark · C4 | AUROC: 88.03 -> 87.09 | Hurt |
| use WaterMod instead of MPAC | multi-bit watermark · GSM8K | accuracy: 31.77 -> 40.33; 26.94% | Improved |
| use WaterMod instead of MPAC | multi-bit watermark · MBPP+ | AUROC: 48.40 -> 98.29; 103.07% | Improved |
| use WaterMod instead of MPAC | multi-bit watermark · MBPP+ | pass@1: 20.60 -> 26.20; 27.18% | Improved |
| use WaterMod instead of MPAC | multi-bit watermark · C4 | perplexity: 10.88 -> 10.87 | No clear effect |
| apply ChatGPT paraphrase attack to WaterMod text | WaterMod · GSM8K | AUROC: 100.00 -> 99.95 | No clear effect |
| apply ChatGPT paraphrase attack to WaterMod text | WaterMod · GSM8K | mean z-score: 14.89 -> 9.95 | Hurt |
Closest papers by meaning
- Break the Trade-off Between Watermark Strength and Speculative Sampling Efficiency for Language Models (ICLR 2026 · similarity 0.89)
- PRO: Enabling Precise and Robust Text Watermark for Open-Source LLMs (NeurIPS 2026 · similarity 0.88)
- Black-Box Detection of Language Model Watermarks (ICLR 2025 · similarity 0.88)
- Theoretically Grounded Framework for LLM Watermarking: A Distribution-Adaptive Approach (NeurIPS 2025 · similarity 0.87)
- Unbiased Watermark for Large Language Models (ICLR 2024 · similarity 0.87)
- GaussMark: A Practical Approach for Structural Watermarking of Language Models (ICML 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
