Chamber LV v1.1.0 RAW TRACK NORM TRACK R5-T SLOPE
Tier-2 Selective Transfer Diagnostic — Encoding Artifact vs Hierarchy Stratification
Central question: Does V2×V4 failure persist after encoding-invariant normalization? If yes → Selective Transfer Theorem. If no → Scale drift artifact.
Decision Rule (preregistered):
Declare A (Encoding Artifact) if: Run B (normalized) meets passFraction at R1=0.70 for all previously failing gates, AND the R1 sweep shows normalized retention curves align with stable families.
Declare B (Selective Transfer) if: V2×V4 still fails at R1=0.70 after normalization (B1), OR has systematically steeper retention erosion slope than stable families (B2). Either condition sufficient.
⚙️ Configuration
V2×V3 (Tier-2 × Tier-1)
V2×V4 (Tier-2 + curvature)
V6×V7 (Tier-2 internal)
G1 (off — no clear scale issue)
G2 — δκ_rel = δκ / (⟨κ⟩ + η)
G3 — B_rel = sharpness / median(B)
G4 — L_rel = L / (⟨L⟩ + η)
R2-T (Overlap): 0.10   R3-T (Interference): 0.05   R4-T (Migration): 0.10
Comma-separated thresholds
📋 Execution Log
[INIT] Chamber LV v1.1.0 initialized.
[INIT] Dual-track (raw + normalized) with R5-T slope index.