The harm of divergence is inherently claim-dependent: the same divergence can be harmless for one mechanistic claim and pernicious for another

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

• How can we produce a principled method for classifying harmful divergence for any mechanistic claim?question0.818
  Identified gap: current work lacks a general method for harmful divergence classification
• No principled method exists for classifying harmful divergence for arbitrary mechanistic claimsconcept0.785
  Explicitly identified limitation: the paper cannot classify perniciousness in general
• When, and to what extent, is it okay for divergences to occur?question0.780
  Second core research question motivating the theoretical analysis in Section 4
• Harmless Divergenceconcept0.780
  Divergences that occur in the behavioral null-space and do not affect functional claims about the model
• Any divergence outside of the null-space of NN layers is potentially pernicious, posing challenges for a complete mechanistic understanding of NNsclaim0.767
  Sobering conclusion about the fundamental challenge posed by divergence for mechanistic interpretability
• Pernicious Divergenceconcept0.762
  Divergences that activate hidden pathways or cause dormant behavioral changes, undermining mechanistic claims
• When it is not okay, how can we prevent divergent representations from occurring?question0.760
  Third core research question motivating the CL loss approach in Section 5
• Minimizing divergence magnitude does not guarantee elimination of hidden pathways; it only reduces the risk surfaceclaim0.753
  Important caveat to the CL loss solution, noting it is a step not a complete fix