concept
active
concept:knowledge-localization

Knowledge Localization

Technique for identifying where specific knowledge is stored in neural network layers via interventions

Neighborhood — ranked by edge-count

Thinkers (1)

thinker
  • Kevin Meng
    studies
    Author of ROME paper on locating and editing factual associations in GPT.

Concepts (1)

concept
  • The field aimed at understanding what neural networks have learned; characterized as pre-paradigmatic in this paper

Related by similarity (8)

cosine ≥ 0.65 · no typed edge

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

  • Chinn et al. showed that tactile target experience promotes earlier mirror self-recognition in infants; noted as a future extension
  • Novel task asking which of 10 sentences received injection, cycling injection through all positions to average out positional bias
  • spatializationconcept0.745
    The translation of semantic values into spatial coordinates and relations.
  • Generalizationconcept0.729
    Ability to apply learned solutions to novel circumstances.
  • The ability to generalize across tasks; lacking in latent methods.
  • Learningconcept0.714
    Inference of parameters encoding contingencies of the world (e.g., likelihood matrix A) at slower timescale than perception.
  • Generalisationconcept0.713
    Ability to respond appropriately to novel situations based on past regularities; fundamental to learning and intelligence.
  • canalizationconcept0.709
    Waddington's concept: developmental buffering that produces a stable phenotype despite genetic/environmental perturbation.