English

Advancing Beyond Identification: Multi-bit Watermark for Large Language Models

Computation and Language 2024-03-21 v3 Artificial Intelligence Cryptography and Security

Abstract

We show the viability of tackling misuses of large language models beyond the identification of machine-generated text. While existing zero-bit watermark methods focus on detection only, some malicious misuses demand tracing the adversary user for counteracting them. To address this, we propose Multi-bit Watermark via Position Allocation, embedding traceable multi-bit information during language model generation. Through allocating tokens onto different parts of the messages, we embed longer messages in high corruption settings without added latency. By independently embedding sub-units of messages, the proposed method outperforms the existing works in terms of robustness and latency. Leveraging the benefits of zero-bit watermarking, our method enables robust extraction of the watermark without any model access, embedding and extraction of long messages (\geq 32-bit) without finetuning, and maintaining text quality, while allowing zero-bit detection all at the same time. Code is released here: https://github.com/bangawayoo/mb-lm-watermarking

Keywords

Cite

@article{arxiv.2308.00221,
  title  = {Advancing Beyond Identification: Multi-bit Watermark for Large Language Models},
  author = {KiYoon Yoo and Wonhyuk Ahn and Nojun Kwak},
  journal= {arXiv preprint arXiv:2308.00221},
  year   = {2024}
}

Comments

NAACL 2024 main. 9 pages and appendix

R2 v1 2026-06-28T11:45:05.395Z