The mechanisms implementing NPI licensing and filler-gap dependencies are learned in discrete stages, not gradually

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

• For both NPI and filler-gap tasks, the model initially learns to move information directly from alternating token to output; intermediate steps are added later in trainingclaim0.845
  Mechanistic interpretation of training dynamics in case studies
• NPI mechanism in pythia-1b moves negation feature through complementiser 'that', auxiliary verb, and main verb across layers before predicting NPI 'any'finding0.756
  Mechanistic finding from CausalGym case study showing multi-step information movement in NPI mechanism
• A family of contrastive learners converges to a representation whose kernel is the pointwise mutual information (PMI) of the underlying eventshypothesis0.752
  Mathematical formalization of what representation models converge to
• The conditions necessary to evolve a new level of individuality are described by the conditions necessary to learn non-decomposable functions (deep model induction)claim0.751
  Core interpretive claim of the paper connecting ETIs to connectionist learning
• Pretraining plays a role analogous to unlabeled experience in humans — building P_prior before semantic binding — explaining why few labeled examples sufficeclaim0.749
  Developmental analogy used to explain sample efficiency under high ρd conditions
• There are fewer representations competent for N tasks than M<N tasks, so training more general models should yield fewer possible solutionshypothesis0.743
  Selective pressure toward convergence via task generality
• Software implementations for all of the models/behaviours presented are common for n = 2, and can be made very efficient for α_i that map many objects onto a much smaller set of object families.claim0.741
  Claim about current practical feasibility and efficiency of 2-way associative implementations.
• The core of all living process is step-by-step adaptation — the modification and evolution which happen gradually in response to information about the degree to which the emerging structure supports the whole.claim0.740
  Definition of the essential mechanism of living structure formation.