Automatic Discovery of Visual Circuits

arXiv:2404.14349v1 Announce Type: new
Abstract: To date, most discoveries of network subcomponents that implement human-interpretable computations in deep vision models have involved close study of single units and large amounts of human labor. We explore scalable methods for extracting the subgraph of a vision model's computational graph that underlies recognition of a specific visual concept. We introduce a new method for identifying these subgraphs: specifying a visual concept using a few examples, and then tracing the interdependence of neuron activations …

abstract arxiv circuits computational concept cs.ai cs.cv discoveries discovery explore graph human labor network recognition scalable study type units vision vision models visual