Vision-Language Models Create Cross-Modal Task Representations
This work provides insights into VLM internal processing for researchers in multimodal AI, though it is incremental in understanding existing models.
The study investigated how autoregressive vision-language models (VLMs) internally represent tasks, finding that they align conceptually equivalent inputs into a shared task vector that is invariant to modality and format, which outperforms full task prompting in cross-modal cases.
Autoregressive vision-language models (VLMs) can handle many tasks within a single model, yet the representations that enable this capability remain opaque. We find that VLMs align conceptually equivalent inputs into a shared task vector, which is invariant to modality (text, image) and format (examples, instruction), and may simplify VLM processing. We measure this alignment via cross-modal transfer -- the ability of a task vector derived in one modality to trigger the correct generation in another -- on a range of tasks and model architectures. Although the task vector is highly compressed, we find that this single vector outperforms prompting the model with the full task information, unique to this cross-modal case. Furthermore, we show that task vectors can be transferred from a base language model to its fine-tuned vision-language counterpart, and that they can be derived solely from instructions without the need for examples. Taken together, our findings shed light on how VLMs internally process task information, and how they map different modalities into common semantic representations. Project page: https://vlm-cross-modal-reps.github.io.