Taking Shortcuts for Categorical VQA Using Super Neurons
This work addresses efficiency and accuracy challenges in VLMs for vision-language tasks, offering a training-free method that is incremental but provides significant practical gains.
The paper tackled the problem of improving Vision Language Models (VLMs) for visually grounded tasks by proposing Super Neurons (SNs), which use scalar activations from shallow layers to create accurate classifiers, achieving a 5.10x speedup while enhancing classification performance.
Sparse Attention Vectors (SAVs) have emerged as an excellent training-free alternative to supervised finetuning or low-rank adaptation to improve the performance of Vision Language Models (VLMs). At their heart, SAVs select a few accurate attention heads for a task of interest and use them as classifiers, rather than relying on the model's prediction. In a similar spirit, we find that directly probing the raw activations of the VLM, in the form of scalar values, is sufficient to yield accurate classifiers on diverse visually grounded downstream tasks. Shifting focus from attention vectors to scalar activations dramatically increases the search space for accurate parameters, allowing us to find more discriminative neurons immediately from the first generated token. We call such activations Super Neurons (SNs). In this probing setting, we discover that enough SNs appear in the shallower layers of the large language model to allow for extreme early exiting from the first layer of the model at the first generated token. Compared to the original network, SNs robustly improve the classification performance while achieving a speedup of up to 5.10x.