Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2403.10020 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915137475051520 |
|---|---|
| author | Luo, Yiyang Lin, Ke Gu, Chao Hou, Jiahui Wen, Lijie Luo, Ping |
| author_facet | Luo, Yiyang Lin, Ke Gu, Chao Hou, Jiahui Wen, Lijie Luo, Ping |
| contents | The proliferation of large language models (LLMs) in generating content raises concerns about text copyright. Watermarking methods, particularly logit-based approaches, embed imperceptible identifiers into text to address these challenges. However, the widespread usage of watermarking across diverse LLMs has led to an inevitable issue known as watermark collision during common tasks, such as paraphrasing or translation. In this paper, we introduce watermark collision as a novel and general philosophy for watermark attacks, aimed at enhancing attack performance on top of any other attacking methods. We also provide a comprehensive demonstration that watermark collision poses a threat to all logit-based watermark algorithms, impacting not only specific attack scenarios but also downstream applications. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2403_10020 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Lost in Overlap: Exploring Logit-based Watermark Collision in LLMs Luo, Yiyang Lin, Ke Gu, Chao Hou, Jiahui Wen, Lijie Luo, Ping Computation and Language Multimedia The proliferation of large language models (LLMs) in generating content raises concerns about text copyright. Watermarking methods, particularly logit-based approaches, embed imperceptible identifiers into text to address these challenges. However, the widespread usage of watermarking across diverse LLMs has led to an inevitable issue known as watermark collision during common tasks, such as paraphrasing or translation. In this paper, we introduce watermark collision as a novel and general philosophy for watermark attacks, aimed at enhancing attack performance on top of any other attacking methods. We also provide a comprehensive demonstration that watermark collision poses a threat to all logit-based watermark algorithms, impacting not only specific attack scenarios but also downstream applications. |
| title | Lost in Overlap: Exploring Logit-based Watermark Collision in LLMs |
| topic | Computation and Language Multimedia |
| url | https://arxiv.org/abs/2403.10020 |