A Multimodal Automated Interpretability Agent

arXiv:2404.14394v1 Announce Type: cross
Abstract: This paper describes MAIA, a Multimodal Automated Interpretability Agent. MAIA is a system that uses neural models to automate neural model understanding tasks like feature interpretation and failure mode discovery. It equips a pre-trained vision-language model with a set of tools that support iterative experimentation on subcomponents of other models to explain their behavior. These include tools commonly used by human interpretability researchers: for synthesizing and editing inputs, computing maximally activating exemplars from real-world datasets, …

abstract agent arxiv automate automated cs.ai cs.cl cs.cv discovery experimentation failure feature interpretability interpretation iterative language language model maia multimodal paper set support tasks tools type understanding vision vision-language