Cross-Modal Projection in Multimodal LLMs Doesn’t Really Project Visual Attributes to Textual Space
Gaurav Verma, Minje Choi, Kartik Sharma, Jamelle Watson-Daniels, Sejoon Oh, Srijan Kumar
Abstract
Multimodal large language models (MLLMs) like LLaVA and GPT-4(V) enable general-purpose conversations about images with the language modality. As off-the-shelf MLLMs may have limited capabilities on images from domains like dermatology and agriculture, they must be fine-tuned to unlock domain-specific applications. The prevalent architecture of current open-source MLLMs comprises two major modules: an image-language (cross-modal) projection network and a large language model. It is desirable to understand the roles of these two modules in modeling domain-specific visual attributes to inform the design of future models and streamline the interpretability efforts on the current models. To this end, via experiments on 4 datasets and under 2 fine-tuning settings, we find that as the MLLM is fine-tuned, it indeed gains domain-specific visual capabilities, but the updates do not lead to the projection extracting relevant domain-specific visual attributes. Our results indicate that the domain-specific visual attributes are modeled by the LLM, even when only the projection is fine-tuned. Through this study, we offer a potential reinterpretation of the role of cross-modal projections in MLLM architectures.- Anthology ID:
- 2024.acl-short.60
- Volume:
- Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)
- Month:
- August
- Year:
- 2024
- Address:
- Bangkok, Thailand
- Editors:
- Lun-Wei Ku, Andre Martins, Vivek Srikumar
- Venue:
- ACL
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 657–664
- Language:
- URL:
- https://aclanthology.org/2024.acl-short.60
- DOI:
- 10.18653/v1/2024.acl-short.60
- Cite (ACL):
- Gaurav Verma, Minje Choi, Kartik Sharma, Jamelle Watson-Daniels, Sejoon Oh, and Srijan Kumar. 2024. Cross-Modal Projection in Multimodal LLMs Doesn’t Really Project Visual Attributes to Textual Space. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 657–664, Bangkok, Thailand. Association for Computational Linguistics.
- Cite (Informal):
- Cross-Modal Projection in Multimodal LLMs Doesn’t Really Project Visual Attributes to Textual Space (Verma et al., ACL 2024)
- PDF:
- https://preview.aclanthology.org/dois-2013-emnlp/2024.acl-short.60.pdf