Watermarking Language Models with Error Correcting Codes
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2024
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866914344908881920 |
|---|---|
| author | Chao, Patrick Sun, Yan Dobriban, Edgar Hassani, Hamed |
| author_facet | Chao, Patrick Sun, Yan Dobriban, Edgar Hassani, Hamed |
| contents | Recent progress in large language models enables the creation of realistic machine-generated content. Watermarking is a promising approach to distinguish machine-generated text from human text, embedding statistical signals in the output that are ideally undetectable to humans. We propose a watermarking framework that encodes such signals through an error correcting code. Our method, termed robust binary code (RBC) watermark, introduces no noticeable degradation in quality. We evaluate our watermark on base and instruction fine-tuned models and find that our watermark is robust to edits, deletions, and translations. We provide an information-theoretic perspective on watermarking, a powerful statistical test for detection and for generating $p$-values, and theoretical guarantees. Our empirical findings suggest our watermark is fast, powerful, and robust, comparing favorably to the state-of-the-art. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_10281 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Watermarking Language Models with Error Correcting Codes Chao, Patrick Sun, Yan Dobriban, Edgar Hassani, Hamed Cryptography and Security Computation and Language Machine Learning Recent progress in large language models enables the creation of realistic machine-generated content. Watermarking is a promising approach to distinguish machine-generated text from human text, embedding statistical signals in the output that are ideally undetectable to humans. We propose a watermarking framework that encodes such signals through an error correcting code. Our method, termed robust binary code (RBC) watermark, introduces no noticeable degradation in quality. We evaluate our watermark on base and instruction fine-tuned models and find that our watermark is robust to edits, deletions, and translations. We provide an information-theoretic perspective on watermarking, a powerful statistical test for detection and for generating $p$-values, and theoretical guarantees. Our empirical findings suggest our watermark is fast, powerful, and robust, comparing favorably to the state-of-the-art. |
| title | Watermarking Language Models with Error Correcting Codes |
| topic | Cryptography and Security Computation and Language Machine Learning |
| url | https://arxiv.org/abs/2406.10281 |