Interpretability Can Be Actionable
For interpretability researchers, this paper provides a conceptual framework to shift focus from method development to actionable outcomes, though it is primarily a position paper without empirical validation.
The paper argues that interpretability research lacks practical impact due to missing evaluation criteria, proposing actionability as a core objective defined by concreteness and validation. It identifies five domains for leverage and presents a framework with outcome-aligned evaluation.
Interpretability aims to explain the behavior of deep neural networks. Despite rapid growth, there is mounting concern that much of this work has not translated into practical impact, raising questions about its relevance and utility. This position paper argues that the central missing ingredient is not new methods, but evaluation criteria: interpretability should be evaluated by actionability--the extent to which insights enable concrete decisions and interventions beyond interpretability research itself. We define actionable interpretability along two dimensions--concreteness and validation--and analyze the barriers currently preventing real-world impact. To address these barriers, we identify five domains where interpretability offers unique leverage and present a framework for actionable interpretability with evaluation criteria aligned with practical outcomes. Our goal is not to downplay exploratory research, but to establish actionability as a core objective of interpretability research.