1
A minimal implementation of LLM output watermarking
用最小代码实现大模型输出水印，保护AI生成内容的溯源利器。
Article URL: https://github.com/berba-q/gpt-watermark Comments URL: https://news.ycombinator.com/item?id=49430302 Points: 1 # Comments: 1
2
Tracing Provenance and Detecting Tampering with Complementary LLM Watermarks
用互补水印机制追踪大模型输出溯源，同时识别篡改痕迹，为AI内容安全提供新思路。
arXiv:2608.12713v1 Announce Type: cross Abstract: Watermarking LLM-generated text is an important task for tracing its provenance. Existing LLM waterm…
3
Efficient Online LLM Watermark Detection via Rao-Blackwellized E-Processes
提出基于Rao-Blackwellized E-Processes的高效在线LLM水印检测方法，大幅降低计算成本，适合AI安全与模型验证场景。
arXiv:2607.21958v1 Announce Type: cross Abstract: As large language models (LLMs) are increasingly deployed, reliable and efficient mechanisms for dis…
4
Making Open-Source Text LLM Watermarks Durable Against Merging
针对开源LLM模型合并后水印失效的痛点，提出新型耐久水印方案，保障AI内容溯源安全。
arXiv:2607.20435v1 Announce Type: cross Abstract: Open-source LLMs (OSMs)arereaching near state-of-the-art performance, prompting prior works to trace…
5
Marking the Wrong Symptoms: Evaluating LLM Watermarks in Medical Texts
从医学文本切入，揭示LLM水印技术的缺陷与挑战，直击当前AI安全盲区。
arXiv:2607.20462v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly integrated into clinical workflows, stressing the need f…
6
ChainMark: Model-Free LLM Watermarking with Closed-Form Calibration
无需访问模型内部，用闭式校准实现LLM文本水印，防篡改可溯源。
arXiv:2607.18445v1 Announce Type: cross Abstract: Regulatory regimes such as the EU AI Act mandate machine-readable marking of synthetic text, but exi…
7
Beyond Heuristic Tuning: Power-Calibrated LLM Watermarking
ICML 2026接收：为LLM水印引入功率校准，告别传统启发式调优。
arXiv:2607.05694v1 Announce Type: cross Abstract: Logit-based watermarking is a widely used mechanism for identifying LLM generated content, yet its e…
8
CORE-BREW: LLR-Based Soft Decoding for Robust Multi-Bit LLM Watermarking
突破性LLM水印方案CORE-BREW采用软解码，提升多比特水印的鲁棒性与误报控制
arXiv:2606.24163v1 Announce Type: cross Abstract: Reliable provenance for LLM outputs requires multi-bit watermarks that remain robust under editing w…
9
Linear Ensembles Wash Away Watermarks: On the Fragility of Distributional Perturbations in LLMs
揭示LLM水印在多重模型访问下的致命缺陷：独立扰动轻松被线性集成抹除，对AI安全与版权保护提出新挑战。
arXiv:2605.30501v1 Announce Type: new Abstract: Watermarking embeds statistical signatures in AI-generated text for detection and attribution. We reve…
10
Blind PRNG Hijacking: An Undetectable Integrity-Preserving Attack Against LLM Watermarking
首次揭示LLM水印的PRNG信任假设漏洞，提出不可检测的完整性破坏攻击，颠覆现有安全认知。
arXiv:2605.28632v1 Announce Type: cross Abstract: Cryptographic watermarking is a leading defense for attributing text generated by large language mod…
11
Robust LLM Watermarking with Minimal Semantic Distortion for IP Protection
最新研究提出鲁棒LLM水印方法，在保护知识产权的同时将语义失真降至最低，平衡了安全性与生成质量。
arXiv:2605.23175v1 Announce Type: cross Abstract: Proprietary large language models (LLMs) face risks of intellectual property (IP) violation, as adve…