Concentration Parameters

Sufficient statistics of Dirichlet priors over likelihood; accumulate as experience is gained; analogous to synaptic efficacy

Neighborhood — ranked by edge-count

concept

Synaptic Homeostasis
analogous_to
Biological mechanism of sleep-dependent synapse elimination; analogous to Bayesian model reduction removing redundant parameters
Dirichlet Distribution
associated_with
Conjugate prior for categorical variables; used for beliefs about likelihood matrix A.
Likelihood Array (A matrix)
associated_with
High-dimensional array mapping hidden states to outcomes; key learned quantity in the paradigm

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.

Temperature / Precision Parameterconcept0.727
The inverse precision (γ = 1/β) that scales the influence of expected free energy on policy selection, controlling exploration-exploitation balance.
Diversity Parameterconcept0.726
Gold standard value (e.g., nucleus sampling p-value) used as ground truth for evaluating diversity metrics
What Parameters Determine Where Organisms Fall On Thequestion0.722
ColSum Concentrationmethod0.721
Primary metric for identifying stages of inference via normalized entropy of column sums of attention matrices
Safety-Critical Parametersconcept0.714
The extremely sparse (~3%) set of parameters governing safety alignment, consistent with its ease of disruption by fine-tuning
Task-specific parameterconcept0.703
Parameter specific to each task, e.g., task head.
Precision Parameter (γ)concept0.693
Core mechanism across all three levels; inverse variance controlling how much the system trusts evidence at each hierarchical level.
Accuracy Criterionconcept0.692
Criterion requiring that model's description of internal state be accurate, distinguishing genuine introspection from confabulation.