aiqbench analysis
This research introduces a novel binomial encoding scheme for LLM watermarks, enabling the embedding of complex payloads like user IDs and timestamps with superior accuracy and robustness compared to existing methods. By dynamically redistributing encoding pressure, the proposed stateful encoder significantly improves performance in high-distortion regimes where previous techniques struggle. The work also establishes per-bit confidence scoring as a critical metric for evaluating practical watermark reliability.
Every Bit, Everywhere, All at Once: A Binomial Multibit LLM Watermark
arXiv:2605.11653v2 Announce Type: replace-cross Abstract: With LLM watermarking already being deployed commercially, practical applications increasingly require multibit watermarks that encode more complex payloads, such as user IDs or timestamps, into the generated text. In this work, we propose a fundamentally new approach for multibit watermarking: introducing binomial encoding to directly encode every bit of the payload at every token position. We complement our approach with a stateful encoder that during generation dynamically redirects encoding pressure toward underencoded bits. Our evaluation against 8 baselines on up to 64-bit payloads shows that our scheme achieves superior message accuracy and robustness, with the gap to baseline methods widening in more relevant settings (i.e., large payloads and low-distortion regimes). At the same time, we challenge prior works' evaluation metrics, highlighting their lack of practical insights, and introduce per-bit confidence scoring as a practically relevant metric for evaluating multibit LLM watermarks.
Abstract:With LLM watermarking already being deployed commercially, practical applications increasingly require multibit watermarks that encode more complex payloads, such as user IDs or timestamps, into the generated text. In this work, we propose a fundamentally new approach for multibit watermarking: introducing binomial encoding to directly encode every bit of the payload at every token position. We complement our approach with a stateful encoder that during generation dynamically redirects encoding pressure toward underencoded bits. Our evaluation against 8 baselines on up to 64-bit payloads shows that our scheme achieves superior message accuracy and robustness, with the gap to baseline methods widening in more relevant settings (i.e., large payloads and low-distortion regimes). At the same time, we challenge prior works' evaluation metrics, highlighting their lack of practical insights, and introduce per-bit confidence scoring as a practically relevant metric for evaluating multibit LLM watermarks.
Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
arXiv:2605.11653 [cs.CR]
(or arXiv:2605.11653v2 [cs.CR] for this version)
https://doi.org/10.48550/arXiv.2605.11653 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 237 words
Related coverage
textGrain watermarking is harder to detect: can sum-tests still do it optimally?
arXiv · Oct 14, 2026
Arvind Gupta : India Rise as a Digital Power • Stimson Center
stimson.org · Aug 1, 2026
Lattice-Based Cryptography Explained
Quantum Zeitgeist · Aug 15, 2026
5 Useful Python Scripts to Automate Boring PDF Tasks
KDnuggets · Jun 10, 2026