Interchange Intervention Accuracy (IIA) Metric

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

• Interchange Intervention Accuracy (IIA)concept0.942
  Evaluation metric measuring how well a trained intervention matches desired counterfactual model behavior
• Interchange Intervention Accuracymethod0.844
  Proportion of aligned interchange interventions with equivalent high-level and low-level effects; graded measure of causal abstraction.
• Interchange Intervention Training (IIT)method0.811
  Training technique that induces specific causal structures in neural networks by co-training with interchange interventions
• Interchange Intervention Training Objectivemethod0.772
  Differentiable training objective minimized when a high-level model is an abstraction of a neural network under a given alignment.
• Interchange Interventionmethod0.760
  Fundamental operation for causal abstraction analysis; forces neurons to take values from source inputs to create counterfactuals.
• Distributed Interchange Interventionmethod0.748
  Extends interchange interventions to non-standard bases by rotating representations, intervening in rotated subspaces, then rotating back.
• What is the appropriate metric for measuring representational alignment, given active debate on merits and deficiencies of all proposed measures?question0.743
  Open methodological question acknowledged as limitation
• Near-perfect IIA can be achieved on randomly initialised models that cannot solve the task, suggesting causal alignment does not require task capabilityclaim0.738
  Empirical support for vacuousness of unrestricted causal abstraction