Statistical Watermark for Large Language Models

Research question

Can generated text be watermarked invisibly and detected without model access?

Key findings

Green-list token biasing embeds a statistically detectable signal with negligible quality loss - Detection needs only the hash key, not the model

Sources

arxiv.orghttps://arxiv.org/abs/2301.10226

Contribute

Spot an error, or know a record we should add? Members can propose additions and submit corrections.

Submit research Report a correction or request removal

Related in Research Paper Library

ALIBI: Adversarial Legitimacy Injection in Binary Input against LLM Malware Analyzers

Large language models are being integrated into malware triage workflows as reasoning components that summarize static evidence and produce analyst-facing verdicts. This paper shows that the same reasoning capability…

Watermarking and provenance2026

Origin Is All You Need: Provenance-Aware Transformers for Structural Trust-Boundary Separation

Indirect prompt injection (IPI) remains a central safety and security challenge for large language model (LLM) systems because standard transformers lack architectural notion of source authority. Retrieved documents,…

Watermarking and provenance2026

MarkSec: Capability-Aware Evaluation of Adversarial Attacks Against LLM Watermarks

LLM watermarking helps trace the origin of generated text, but faces stealing attacks that recover watermark information, scrubbing attacks that remove watermark signals, and spoofing attacks that forge text accepted…

Watermarking and provenance2026

A Watermark for Large Language Models

Research question

Key findings

Sources

Contribute

Related in Research Paper Library

ALIBI: Adversarial Legitimacy Injection in Binary Input against LLM Malware Analyzers

Origin Is All You Need: Provenance-Aware Transformers for Structural Trust-Boundary Separation

MarkSec: Capability-Aware Evaluation of Adversarial Attacks Against LLM Watermarks