VLM-Guard: Safeguarding Vision-Language Models via Fulfilling Safety Alignment Gap
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866929715913162752 |
|---|---|
| author | Liu, Qin Wang, Fei Xiao, Chaowei Chen, Muhao |
| author_facet | Liu, Qin Wang, Fei Xiao, Chaowei Chen, Muhao |
| contents | The emergence of vision language models (VLMs) comes with increased safety concerns, as the incorporation of multiple modalities heightens vulnerability to attacks. Although VLMs can be built upon LLMs that have textual safety alignment, it is easily undermined when the vision modality is integrated. We attribute this safety challenge to the modality gap, a separation of image and text in the shared representation space, which blurs the distinction between harmful and harmless queries that is evident in LLMs but weakened in VLMs. To avoid safety decay and fulfill the safety alignment gap, we propose VLM-Guard, an inference-time intervention strategy that leverages the LLM component of a VLM as supervision for the safety alignment of the VLM. VLM-Guard projects the representations of VLM into the subspace that is orthogonal to the safety steering direction that is extracted from the safety-aligned LLM. Experimental results on three malicious instruction settings show the effectiveness of VLM-Guard in safeguarding VLM and fulfilling the safety alignment gap between VLM and its LLM component. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2502_10486 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | VLM-Guard: Safeguarding Vision-Language Models via Fulfilling Safety Alignment Gap Liu, Qin Wang, Fei Xiao, Chaowei Chen, Muhao Cryptography and Security Artificial Intelligence Computer Vision and Pattern Recognition The emergence of vision language models (VLMs) comes with increased safety concerns, as the incorporation of multiple modalities heightens vulnerability to attacks. Although VLMs can be built upon LLMs that have textual safety alignment, it is easily undermined when the vision modality is integrated. We attribute this safety challenge to the modality gap, a separation of image and text in the shared representation space, which blurs the distinction between harmful and harmless queries that is evident in LLMs but weakened in VLMs. To avoid safety decay and fulfill the safety alignment gap, we propose VLM-Guard, an inference-time intervention strategy that leverages the LLM component of a VLM as supervision for the safety alignment of the VLM. VLM-Guard projects the representations of VLM into the subspace that is orthogonal to the safety steering direction that is extracted from the safety-aligned LLM. Experimental results on three malicious instruction settings show the effectiveness of VLM-Guard in safeguarding VLM and fulfilling the safety alignment gap between VLM and its LLM component. |
| title | VLM-Guard: Safeguarding Vision-Language Models via Fulfilling Safety Alignment Gap |
| topic | Cryptography and Security Artificial Intelligence Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2502.10486 |