Disappearing Ink: Obfuscation Breaks N-gram Code Watermarks in Theory and Practice

arXiv:2507.05512 · cs.CR, cs.AI · Submitted 2025-07-07 · Read on arXiv

cs.CR, cs.AI

Submitted: 2025-07-07

Updated: 2026-08-30

Code: https://github.com/mishoo/UglifyJS

Terminology

• AUROC — 62× in this paper · explained in 2 episodes
• Robustness — 36× in this paper · explained in 5 episodes
• Consistency — 29× in this paper · explained in 3 episodes
• Benchmark — 24× in this paper · explained in 2 episodes
• Generative Model — 7× in this paper · explained in 2 episodes
• Verification — 6× in this paper · explained in 2 episodes
• Adversarial Examples — 4× in this paper · explained in 2 episodes
• Adam — 3× in this paper · explained in 2 episodes
• Generative Models — 3× in this paper · explained in 2 episodes
• Large Language Models (LLMs) — 3× in this paper · explained in 6 episodes
• Benchmarking — 2× in this paper · explained in 3 episodes
• Core — 2× in this paper · explained in 2 episodes
• Ablation — 1× in this paper · explained in 1 episode
• Accuracy — 1× in this paper · explained in 1 episode
• Adversarial scenarios — 1× in this paper · explained in 1 episode
• Area Under the Receiver Operating Characteristic Curve — 1× in this paper
• Generalization — 1× in this paper · explained in 13 episodes
• Provenance — 1× in this paper · explained in 4 episodes
• Transformer — 1× in this paper · explained in 2 episodes
• Verifier — 1× in this paper · explained in 2 episodes

Sources

• Evaluating Large Language Models Trained on Code
• DeepSeek-Coder: When the Large Language Model Meets Programming -- The Rise of Code Intelligence
• GPT-4o System Card
• RankGen: Improving Text Generation with Large Ranking Models
• Adaptive Text Watermark for Large Language Models
• MCGMark: An Encodable and Robust Online Watermark for Tracing LLM-Generated Malicious Code
• Can AI-Generated Text be Reliably Detected?
• LLaMA: Open and Efficient Foundation Language Models
• SoK: Watermarking for AI-Generated Content

Related papers

• SoK: AI-Augmented Binary Reversing
• Relaxed Sender Anonymity for CBDC Interbank Settlement: A Zero-Knowledge Approach on Permissioned EVM
• Calibration-Family Overfit: Why Trusted Sabotage Monitors Don't Transfer Across Lineages
• Efficient Fuzzy PSI under One-Sided Assumptions
• Sealing the Audit-Runtime Gap for LLM Skills
• Token Composition: A Graph Based on EVM Logs