AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

In-Context Watermarks for Large Language Models

Yepeng Liu, Xuandong Zhao, Christopher Kruegel, Dawn Song, Yuheng Bu

ICLR 2026 · Large language models › Language, safety and interpretability · approach: Text watermarking · 15 citations (OpenAlex)

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 changedWhereMeasuredEffect
use stronger LLM (GPT-o3-mini instead of GPT-4o-mini) (ablation)Initials ICW · ELI5 (DTS)ROC-AUC: 0.572 -> 0.999Improved
use stronger LLM (GPT-o3-mini instead of GPT-4o-mini) (ablation)Acrostics ICW · ELI5 (DTS)ROC-AUC: 0.590 -> 1.000Improved
use stronger LLM (GPT-o3-mini instead of GPT-4o-mini) (ablation)Lexical ICW · ELI5 (DTS)ROC-AUC: 0.910 -> 0.995Improved
use stronger LLM (GPT-o3-mini instead of GPT-4o-mini) (ablation)Unicode ICW · ELI5 (DTS)ROC-AUC: 1.000 -> 1.000No clear effect
long-context injection via paper (IPI) instead of system prompt (DTS) (ablation)Acrostics ICW with GPT-o3-mini · ICLR papersROC-AUC: 1.000 -> 0.997No clear effect
paraphrase watermarked text with ChatGPT (ablation)Unicode ICW · ELI5 (DTS)ROC-AUC: 1.000 -> 0.500Hurt
30% synonym replacement attack (ablation)Lexical ICW · ELI5 (DTS)ROC-AUC: 0.995 -> 0.758Hurt
30% deletion attack (ablation)Initials ICW · ELI5 (DTS)ROC-AUC: 0.999 -> 0.998No clear effect
paraphrase attack (ablation)Acrostics ICW · ELI5 (DTS)ROC-AUC: 1.000 -> 0.922Hurt
increase output length from 300 to 1000 words (ablation)Initials ICW · ELI5 (DTS)ROC-AUC: 0.999 -> 0.998No clear effect
ICW instead of PostMark (text quality)GPT-o3-mini · ELI5 text qualityLLM-judge overall: 2.997 -> 4.813Improved
Initials ICW instruction on text quality (ablation)GPT-o3-mini · ELI5 text qualityLLM-judge overall: 4.992 -> 4.282Hurt

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy