Causal importance

A measure of whether a subcomponent is necessary to reproduce model behavior on a specific prompt, predicted by the causal importance network.

Neighborhood — ranked by edge-count

concept

causal importance network
implementsrelated_to
Auxiliary model trained alongside VPD to predict which subcomponents are causally important for each prompt, enabling mechanistic isolation of components.

cosine ≥ 0.65 · no typed edge

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

Causal powerconcept0.798
The ability of an agent to be a driver of subsequent events; a hallmark of cognition that causal emergence quantifies.
Causal Invarianceconcept0.795
Property that causal mechanisms remain stable across environments; desirable for OOD.
Causal Mechanismconcept0.795
Function determining the value of a variable based on its causal parents in an acyclic causal model.
Causal abstractionconcept0.794
A framework the paper uses alongside feature geometry to deepen mechanistic understanding of LMs
Causal Primitive: Necessityconcept0.793
Probability that effect would not occur without cause; counterfactual primitive.
causal regularitiesconcept0.782
The structural-realist grounding for self-evidencing after the bounded self is relinquished.
Causal Influence Diagramsframework0.777
Framework informing path-specific objectives by identifying causal chains leading to risky behaviors
Causal Tracingconcept0.774
Mechanistic interpretability technique for locating factual associations, mentioned as future work direction.