Minimizing divergence magnitude does not guarantee elimination of hidden pathways; it only reduces the risk surface

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

• Off-manifold divergences can activate hidden pathways that produce misleadingly confirmatory behavior while the true mechanism is never exercisedclaim0.824
  Core claim about why pernicious divergence undermines mechanistic conclusions
• Patching h[1] with a divergent representation can activate distinct, hidden pathways that result in misleadingly confirmatory behavior and/or undetected behavior.quote0.780
  Load-bearing description of the core pernicious divergence mechanism illustrated in Figure 1
• Any system that persists must minimise surprisal, thereby gathering evidence for its own generative model.quote0.764
  Opening sentence defining self-evidencing.
• Larger hidden representations create more random structure that DAS can search through, allowing manipulation of counterfactual behavior even in randomly initialized networkshypothesis0.758
  Tested in Section 4.4 calibration experiment; confirmed by findings.
• probably helps not only with faithful reconstruction but also creates interference patterns that encode nuanced information about the deltas and convergences between states.quote0.757
  Key quote connecting path redundancy to interferometric information encoding.
• Any system that persists must minimise surprisal, thereby gathering evidence for its own generative model, a process known as self-evidencing.claim0.756
  Foundational claim of the paper, defining self-evidencing.
• The harm of divergence is inherently claim-dependent: the same divergence can be harmless for one mechanistic claim and pernicious for anotherclaim0.753
  Important nuance that prevents a universal classification of divergence as always good or bad
• Any divergence outside of the null-space of NN layers is potentially pernicious, posing challenges for a complete mechanistic understanding of NNsclaim0.752
  Sobering conclusion about the fundamental challenge posed by divergence for mechanistic interpretability