Phase 1 — Layer Specialisation

simulated preview

Each UCA adapter layer trains independently against its own objective. The base backbone is frozen.

L2 LimbicL3 HippocampalL4 AssociationL5 ExecutiveL6 Metacognitive

total loss

—

task loss

—

calib loss

—

mem loss

—

exec ent.

—

confidence

—

mem hits

—

step

0/600

loss curves

layer status

Sensory

frozen

Limbic

frozen

Hippocampal

frozen

Association

frozen

Executive

frozen

Metacognitive

frozen

training log

Configure and click start training to begin.