Interpretability features converge across different model architectures, revealing structural similarities.

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

• Training identical architectures on the same data with different objective functions should produce systematically different internal evaluative representations, detectable through interpretability tools, even when final task performance is matchedhypothesis0.814
  Second falsifiable prediction linking objective function structure to valence profile
• Different models cannot converge to the same representation if they have access to fundamentally different information; convergence is capped by mutual information between input signalsclaim0.812
  Key limitation of the PRH for non-bijective observations
• How do representations differ or converge between architectures, tasks, and modalities?question0.803
  Broader research question MAS is positioned to address, citing multiple recent works.
• Lack of rigorous cross-model comparison demonstrating that specific named features (not just correlated ones) form across architecturesquestion0.797
  Explicitly identified research gap: anecdotal evidence exists but rigorous characterization is absent
• The field of interpretability has focused mainly on understanding model activations, not the computations themselvesclaim0.791
  Motivation for VPD's parameter-focused approach.
• Bigger models are more likely to converge to a shared representation than smaller modelshypothesis0.788
  Selective pressure toward convergence via model capacity
• There is a growing similarity in how datapoints are represented in different neural network models, spanning different architectures, training objectives, and data modalitiesclaim0.782
  Primary empirical claim of the paper
• Diverse computer vision models trained on visual recognition tasks converge to remarkably similar internal feature representations regardless of architecture, training procedure, or implementation details, closely matching the organization of animal visual cortexfinding0.779
  Empirical evidence for the universality hypothesis cited as supporting the possibility of convergent consciousness-like solutions