English(EN) MarkTune: Improving the Quality-Detectability Trade-off in Model-Embedded LLM Watermarking

新的MarkTune框架将大语言模型水印直接嵌入模型权重

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-23 04:00

研究人员开发了MarkTune，一个将水印直接嵌入大语言模型权重的新框架。这种方法避免了其他方法的推理时间开销和限制，使其适用于开放权重模型。MarkTune通过将模型正则化到高质量参考分布，从而解决了质量-可检测性权衡问题，在保持生成质量的同时，能够抵御各种攻击并保持强大的检测能力。 AI

影响提供了一种将水印嵌入大语言模型的方法，且不影响生成质量或推理速度，这对于开放权重模型至关重要。

排序理由关于大语言模型水印新方法的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 4 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 关于大语言模型水印新方法的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: paper, model release; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: Same-day; Cluster formed today. Ranking reflects the current source set at time of score.

完整方法见我们的编辑标准。

arXiv cs.LG TIER_1 English(EN) · Yizhou Zhao, Zhiwei Steven Wu, Adam Block · 2026-10-23 04:00

MarkTune: Improving the Quality-Detectability Trade-off in Model-Embedded LLM Watermarking

arXiv:2512.04044v2 Announce Type: replace Abstract: Language model watermarking schemes fall into two broad categories: inference-time methods, which modify the decoding process at generation time, and model-embedded methods, which encode a secret signal directly into the model w…