CodeIP: A Grammar-Guided Multi-Bit Watermark for Large Language Models of Code
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866929650206244864 |
|---|---|
| author | Guan, Batu Wan, Yao Bi, Zhangqian Wang, Zheng Zhang, Hongyu Zhou, Pan Sun, Lichao |
| author_facet | Guan, Batu Wan, Yao Bi, Zhangqian Wang, Zheng Zhang, Hongyu Zhou, Pan Sun, Lichao |
| contents | Large Language Models (LLMs) have achieved remarkable progress in code generation. It now becomes crucial to identify whether the code is AI-generated and to determine the specific model used, particularly for purposes such as protecting Intellectual Property (IP) in industry and preventing cheating in programming exercises. To this end, several attempts have been made to insert watermarks into machine-generated code. However, existing approaches are limited to inserting only a single bit of information. In this paper, we introduce CodeIP, a novel multi-bit watermarking technique that inserts additional information to preserve crucial provenance details, such as the vendor ID of an LLM, thereby safeguarding the IPs of LLMs in code generation. Furthermore, to ensure the syntactical correctness of the generated code, we propose constraining the sampling process for predicting the next token by training a type predictor. Experiments conducted on a real-world dataset across five programming languages demonstrate the effectiveness of CodeIP in watermarking LLMs for code generation while maintaining the syntactical correctness of code. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2404_15639 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | CodeIP: A Grammar-Guided Multi-Bit Watermark for Large Language Models of Code Guan, Batu Wan, Yao Bi, Zhangqian Wang, Zheng Zhang, Hongyu Zhou, Pan Sun, Lichao Computation and Language Large Language Models (LLMs) have achieved remarkable progress in code generation. It now becomes crucial to identify whether the code is AI-generated and to determine the specific model used, particularly for purposes such as protecting Intellectual Property (IP) in industry and preventing cheating in programming exercises. To this end, several attempts have been made to insert watermarks into machine-generated code. However, existing approaches are limited to inserting only a single bit of information. In this paper, we introduce CodeIP, a novel multi-bit watermarking technique that inserts additional information to preserve crucial provenance details, such as the vendor ID of an LLM, thereby safeguarding the IPs of LLMs in code generation. Furthermore, to ensure the syntactical correctness of the generated code, we propose constraining the sampling process for predicting the next token by training a type predictor. Experiments conducted on a real-world dataset across five programming languages demonstrate the effectiveness of CodeIP in watermarking LLMs for code generation while maintaining the syntactical correctness of code. |
| title | CodeIP: A Grammar-Guided Multi-Bit Watermark for Large Language Models of Code |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2404.15639 |