How Much Do Circuits Tell Us? Measuring the Consistency and Specificity of Language Model Circuits
For mechanistic interpretability researchers, it reveals that discovered circuits are not task-specific, questioning their utility for targeted understanding and intervention.
The paper measures circuit reuse, consistency, and specificity in language model circuits across six tasks and seven models, finding high within-task reuse but low task-specificity: ablating one task's circuit damages another task's performance almost as much as its own circuit. This suggests circuits at the component level lack targeted interpretability.
The circuits framework in mechanistic interpretability aims to identify causally important sparse subgraphs of model components, typically evaluated by measuring necessity and sufficiency. We measure circuit reuse, the proportion of components shared across per-example circuits within a task, and investigate two less-studied properties of this: consistency, the recurrence of components within a task, and specificity, their uniqueness to a task. Using edge attribution patching across six tasks and seven models, we find that within-task reuse is high and that shared components are necessary for task performance, with ablations causing up to $\sim$100% relative accuracy drops. However, circuits turn out not to be task-specific: ablating one task's circuit damages another task's performance about as much as that task's own circuit does. We discover that this is due to substantial overlap between circuits across tasks, which are causally important for performance. Some circuits do contain a smaller set of task-specific components, but these account for only a modest portion of circuit performance. Overall, our findings suggest that while circuit discovery at the level of attention heads and MLP layers identifies important components, their lack of task-specificity raises questions about the degree to which circuits can support targeted understanding and intervention on model behavior.