concept
active
concept:neural-codeNeural code
The model's parameters considered as the actual 'code' implementing its algorithms, as opposed to human-written code.
Neighborhood — ranked by edge-count
Claims (1)
claim
- Language models implement algorithms humans have tried and failed to write by hand for decadesassociated_withOpening interpretive claim about the remarkable nature of language models.
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- Load-bearing framing of the core interpretability problem: neural networks encode algorithms in parameter matrices rather than human-readable code.
- Cognition in nervous systems, used as a modelling target
- Brain-based physical implementations of consciousness-related functions, assumed by many ToCs to be exclusive.
- Prior framework combining cellular automata with deep learning, extended by this work
- Michael Johnson's prior work on how neural networks (and brains) can be 'annealed' to find optimal states.