The Security Threat of Compressed Projectors in Large Vision-Language Models
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2025
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866908576479444992 |
|---|---|
| author | Zhang, Yudong Xie, Ruobing Sun, Xingwu Chen, Jiansheng Kang, Zhanhui Wang, Di Wang, Yu |
| author_facet | Zhang, Yudong Xie, Ruobing Sun, Xingwu Chen, Jiansheng Kang, Zhanhui Wang, Di Wang, Yu |
| contents | The choice of a suitable visual language projector (VLP) is critical to the successful training of large visual language models (LVLMs). Mainstream VLPs can be broadly categorized into compressed and uncompressed projectors, and each offers distinct advantages in performance and computational efficiency. However, their security implications have not been thoroughly examined. Our comprehensive evaluation reveals significant differences in their security profiles: compressed projectors exhibit substantial vulnerabilities, allowing adversaries to successfully compromise LVLMs even with minimal knowledge of structure information. In stark contrast, uncompressed projectors demonstrate robust security properties and do not introduce additional vulnerabilities. These findings provide critical guidance for researchers in selecting optimal VLPs that enhance the security and reliability of visual language models. The code is available at https://github.com/btzyd/TCP. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_00534 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | The Security Threat of Compressed Projectors in Large Vision-Language Models Zhang, Yudong Xie, Ruobing Sun, Xingwu Chen, Jiansheng Kang, Zhanhui Wang, Di Wang, Yu Cryptography and Security Artificial Intelligence The choice of a suitable visual language projector (VLP) is critical to the successful training of large visual language models (LVLMs). Mainstream VLPs can be broadly categorized into compressed and uncompressed projectors, and each offers distinct advantages in performance and computational efficiency. However, their security implications have not been thoroughly examined. Our comprehensive evaluation reveals significant differences in their security profiles: compressed projectors exhibit substantial vulnerabilities, allowing adversaries to successfully compromise LVLMs even with minimal knowledge of structure information. In stark contrast, uncompressed projectors demonstrate robust security properties and do not introduce additional vulnerabilities. These findings provide critical guidance for researchers in selecting optimal VLPs that enhance the security and reliability of visual language models. The code is available at https://github.com/btzyd/TCP. |
| title | The Security Threat of Compressed Projectors in Large Vision-Language Models |
| topic | Cryptography and Security Artificial Intelligence |
| url | https://arxiv.org/abs/2506.00534 |