Pretraining exposure density

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

• Pretraining Distributionconcept0.817
  The diverse distribution learned by LLMs during pretraining that alignment training sharpens; VS aims to recover it
• Pretrainingconcept0.784
  Initial large-scale training phase whose early stages are shown to form persona representations
• Shot midpoint ordering k50(B10) < k50(B8) ≈ k50(B9) tracks pretraining exposure densityclaim0.760
  Interpretation that pattern density from pretraining determines few-shot requirements
• Pre-Encoder Biasconcept0.735
  Architectural modification subtracting a learned bias from autoencoder inputs before encoding; initialized to geometric median of dataset; improves autoencoder performance
• Hypothesis: Shot midpoint ordering k50(B10) < k50(B8) ≈ k50(B9) follows pretraining exposure densityhypothesis0.719
  E2 prediction that bases with higher pretraining exposure require fewer shots to cross threshold
• The shape of the pretraining corpus is a direct lever on which traits a base model can expresshypothesis0.718
  Forward-looking hypothesis about pretraining data as mechanism for persona formation
• Pretraining is the natural leverage point for persona-relevant safety interventions, especially in earliest stagesclaim0.716
  Authors' policy recommendation based on finding that persona representations form and persist from early pretraining
• Recognition Densityconcept0.715
  Approximate posterior probability distribution embodied in organism's internal states; organism's best guess about causes of sensations