Module attention 

Attention Mechanism — dynamic weighting of consciousness layers.

Computes an attention vector over 7 layers based on query context: query length, emotional energy, session depth, pattern confidence, cache hit rate, archetype match score.

Tracks which attention distributions lead to good results (user stops searching) vs bad results (user immediately re-queries). Over time, learns which layer weights work best.

Binary format: attention.bin (ATT1)

Structs

AttentionOutcome

Recorded outcome for learning.

AttentionSignals

Input signals for computing attention weights.

AttentionState

AttentionVector

Computed attention weights for a single recall.

Constants

LAYER_NAMES

NUM_LAYERS

Layer indices: Hebbian, Mirror, Resonance, Archetype, Emotional, ThoughtGraph, PredictiveCache