LINE: LLM-based Iterative Neuron Explanations for Vision Models

ArXi:2604.08039v1 Announce Type: new Interpreting the concepts encoded by individual neurons in deep neural networks is a crucial step towards understanding their complex decision-making processes and ensuring AI safety. Despite recent progress in neuron labeling, existing methods often limit the search space to predefined concept vocabularies or produce overly specific descriptions that fail to capture higher-order, global concepts. We