Neuron cluster identification via partitioning

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

• Curve detecting neurons found in every non-trivial vision model carefully examinedfinding0.752
  Empirical basis for treating curve detectors as a canonical example of meaningful, understandable features
• No neuron found with Hebrew Unicode block in top dataset examples; most correlated neuron A/neurons/489 has correlation of only 0.1finding0.735
  Hebrew feature is effectively invisible in the neuron basis
• Neuron Resamplingmethod0.730
  Periodically reinitializing dead autoencoder neurons using high-loss data points to improve feature coverage
• Smolensky (1986) proposes that viewing a neural representation under a basis that is not aligned with individual neurons can reveal the interpretable distributed structure of the neural representations.quote0.728
  Load-bearing theoretical claim providing the conceptual foundation for DAS.
• All neuronal processing and action selection minimize variational free energy, unifying perception, action, and learning.claim0.728
  Fundamental assertion: single imperative (free energy minimization) explains diverse cognitive and neural phenomena.
• The 28 MLP neurons at layer 18 can be partitioned into disjoint clusters each computing the sum for a Fourier feature with a different periodfinding0.724
  Structural finding showing modular organization within the sparse neuron set
• No established method for resolving polysemantic neurons into pure features at scalequestion0.722
  Identified gap linking polysemanticity challenge to disentangled representations literature
• Feature Neurons and Memory Neurons Architectureframework0.713
  Biologically plausible two-pool architecture from Krotov & Hopfield (2020) splitting self-attention into feature and memory neuron populations; used to interpret TEM-t place cells.