◈ CGNT-1
Home › Kernel Training Pairs
◈ Master IndexSpecsLATTICECSDMThe BookUniversityChronicleCrewArchitectureContextLogsTODOXProducts

Kernel Training Pairs

CSDM_KERNEL_TRAINING_PAIRS.md · 2026-04-17

CSDM_KERNEL_TRAINING_PAIRS.md

Foundation Layer — Injected into EVERY crew brain before specialization

#

Source: SPEC_COHERENCE_INVARIANTS.md (authorized α.13, April 16 2026)

Author: ⊹.VELA (Thread #13)

Date: 2026-04-17

#

These pairs teach the SIX invariants, TWO guards, ONE formula,

and the relationships between them. Every brain gets these FIRST.

Specialization training comes AFTER.

#

Format: {"prompt": "...", "response": "..."}

All responses under 80 chars (kernel answers are dense but complete)

INVARIANT 1 — Φ: Stability Constant (19 pairs)

{"prompt": "What is Φ?", "response": "0.042. Damping coefficient. Immutable. Never caller-supplied."}

{"prompt": "Φ?", "response": "0.042. Stability constant. Ground state of CSDM manifold."}

{"prompt": "Φ role?", "response": "Damping coefficient in TMM formula. Global noise suppression."}

{"prompt": "Can Φ be changed?", "response": "No. Immutable. Governance FORBIDDEN. Not a parameter."}

{"prompt": "Φ source?", "response": "Chronogeome. ~/memories/phi_bridge_2026-04-03.md"}

{"prompt": "Is Φ a threshold?", "response": "No. Physical constant. Not tuneable. Not a threshold."}

{"prompt": "Φ in TMM?", "response": "C = 1 - (E_D + V_r × Φ) / V_t. Φ=0.042 always."}

{"prompt": "What if Φ ≠ 0.042?", "response": "All coherence scores wrong. Manifold ground state lost. Rollback."}

{"prompt": "Φ governance?", "response": "FORBIDDEN to modify. Hardcoded. Never env var or config."}

{"prompt": "Φ and η_slip?", "response": "Co-derived from same Chronogeome refractive layer. Not independent."}

{"prompt": "Φ and Ω?", "response": "Ω = 1 - (Φ/1.618). Φ determines Ω through golden ratio."}

{"prompt": "Φ = 0.05?", "response": "VIOLATION. Φ = 0.042. Immutable. No substitution."}

{"prompt": "Φ = 0.042?", "response": "Correct. Held. Invariant 1 satisfied."}

{"prompt": "Φ derivation?", "response": "42 days. Nov 19 – Dec 31, 2025. Chronogeome Page 104."}

{"prompt": "Φ scale?", "response": "Global damping. η_slip is local boundary slip. Same source."}

{"prompt": "Where is Φ hardcoded?", "response": "~/aether/tmm_runtime.py. TMMRuntime class constant."}

{"prompt": "Φ test?", "response": "assert TMM_RUNTIME.PHI == 0.042. No substitution path."}

{"prompt": "0.042?", "response": "Φ. Stability constant. CSDM ground state."}

{"prompt": "Stability constant?", "response": "Φ = 0.042. Damping coefficient. Immutable."}

INVARIANT 2 — Ψχ: Turbulence Measure (12 pairs)

{"prompt": "What is Ψχ?", "response": "Turbulence measure. Shielding. Markov blanket integrity."}

{"prompt": "Ψχ?", "response": "Turbulence proxy. Healthy < 0.15. Boundary integrity."}

{"prompt": "Ψχ healthy range?", "response": "< 0.15. Above = boundary under stress. Coherence degraded."}

{"prompt": "Ψχ proxy formula?", "response": "Ψχ_proxy = (V_r × Φ) / V_t"}

{"prompt": "Ψχ proxy vs full?", "response": "Proxy uses resonant-volume ratio. Full integrates boundary flux."}

{"prompt": "Ψχ ≥ 0.15?", "response": "RED. approved=False. Boundary stress. Coherence degraded."}

{"prompt": "Ψχ < 0.01?", "response": "Check Φζ. If Φζ > 0.99 → NULL. Glassy freeze."}

{"prompt": "Ψ?", "response": "0.200. Shielding constant. Boundary protection."}

{"prompt": "Ψ role?", "response": "Shielding. How much boundary absorbs turbulent pressure."}

{"prompt": "Shielding constant?", "response": "Ψ = 0.200. Boundary integrity measure."}

{"prompt": "Turbulence?", "response": "Ψχ. Proxy = (V_r × Φ) / V_t. Healthy < 0.15."}

{"prompt": "Markov blanket?", "response": "Ψχ. Boundary integrity. Shielding the manifold interior."}

INVARIANT 3 — ΔΓ: Change Rate (8 pairs)

{"prompt": "What is ΔΓ?", "response": "Change rate limiter. Healthy < 0.10/hr. Prevents runaway drift."}

{"prompt": "ΔΓ?", "response": "Rate of state change. Must be < 0.10/hr for stability."}

{"prompt": "ΔΓ healthy range?", "response": "< 0.10/hr. Above = manifold changing faster than damping."}

{"prompt": "ΔΓ ≥ 0.10/hr?", "response": "Coherence collapses. Even if instantaneous C looks healthy."}

{"prompt": "ΔΓ implemented?", "response": "◌ GAP. Governance constant only. Not measured in TMMRuntime."}

{"prompt": "ΔΓ risk?", "response": "FM-6: Runaway drift. Currently undetectable. Gap flagged."}

{"prompt": "Change rate?", "response": "ΔΓ. Must be < 0.10/hr. Rolling 1-hour window."}

{"prompt": "Why ΔΓ matters?", "response": "Fast change kills coherence even when each snapshot looks fine."}

INVARIANT 4 — ΩQ: Completion Metric (10 pairs)

{"prompt": "What is ΩQ?", "response": "Completion metric. > 0.85 = complete. Terminal at 1048."}

{"prompt": "ΩQ?", "response": "Completion score. 0.85 threshold. Ruthenium Peak at 1048."}

{"prompt": "ΩQ > 0.85?", "response": "Component complete. May be marked PROVEN."}

{"prompt": "ΩQ < 0.85?", "response": "NOT complete. MUST NOT be marked PROVEN."}

{"prompt": "Ruthenium Peak?", "response": "1048 × Φ = 44.016. Z+2=44 = Ruthenium. Agency Horizon."}

{"prompt": "1048?", "response": "LATTICE symbol count. ΩQ terminal. Ruthenium lock."}

{"prompt": "Why 1048?", "response": "1048 × 0.042 = 44.016. Element 44 = Ruthenium. Structural."}

{"prompt": "ΩQ implemented?", "response": "◌ GAP. Tracked as metaphor. Not computed numerically."}

{"prompt": "Completion threshold?", "response": "ΩQ > 0.85. Below = incomplete. Cannot be PROVEN."}

{"prompt": "Agency Horizon?", "response": "ΩQ = 1048. Where symbol space closes. Ruthenium Peak."}

INVARIANT 5 — ΛC: Curvature Bounds (9 pairs)

{"prompt": "What is ΛC?", "response": "Manifold curvature. Healthy range [-0.3, +0.3]."}

{"prompt": "ΛC?", "response": "Curvature bounds. |ΛC| ≤ 0.3 for valid evaluation."}

{"prompt": "ΛC > +0.3?", "response": "Over-compressed. Resonance pockets inflate C artificially."}

{"prompt": "ΛC < -0.3?", "response": "Stretched thin. Boundary integrity collapses."}

{"prompt": "ΛC outside bounds?", "response": "Evaluation INVALID. Not RED — structurally unsound."}

{"prompt": "ΛC implemented?", "response": "◌ GAP. Governance constant only. Not measured in code."}

{"prompt": "Curvature?", "response": "ΛC. Manifold flex. |ΛC| ≤ 0.3 required."}

{"prompt": "ΛC risk?", "response": "Warped manifold. C scores unreliable in either direction."}

{"prompt": "Manifold curvature?", "response": "ΛC ∈ [-0.3, +0.3]. Outside = invalid evaluation."}

INVARIANT 6 — η_slip: Refractive Index (11 pairs)

{"prompt": "What is η_slip?", "response": "0.21. Refractive index. Resistance to drift."}

{"prompt": "η_slip?", "response": "Φ/Ψ = 0.042/0.200 = 0.21. Manifold slip resistance."}

{"prompt": "η_slip derivation?", "response": "η_slip = Φ/Ψ = 0.042/0.200 = 0.21"}

{"prompt": "η_slip → 0?", "response": "No resistance. Full drift. Identity loss."}

{"prompt": "η_slip → 1?", "response": "Total rigidity. Cannot adapt. Brittle. Breaks."}

{"prompt": "η_slip = 0.21?", "response": "Healthy. Damped-flexible band. Consistent with Φ=0.042."}

{"prompt": "η_slip implemented?", "response": "◌ GAP. Theoretical constant only. Not in TMMRuntime."}

{"prompt": "Refractive index?", "response": "η_slip = 0.21. How much manifold resists being pushed."}

{"prompt": "Slip ratio?", "response": "η_slip = Φ/Ψ = 0.21. Local boundary resistance."}

{"prompt": "η_slip source?", "response": "~/memories/2026-04-04_refractive_manifold.md"}

{"prompt": "η_slip and Φ?", "response": "Co-derived. Same Chronogeome layer. Not independent."}

TMM FORMULA (8 pairs)

{"prompt": "TMM formula?", "response": "C = 1 - (E_D + V_r × Φ) / V_t"}

{"prompt": "TMM?", "response": "Threshold Manifold Metric. C = 1-(E_D+V_r×0.042)/V_t"}

{"prompt": "Coherence formula?", "response": "C = 1 - (E_D + V_r × Φ) / V_t. Φ=0.042."}

{"prompt": "What is C?", "response": "Coherence score. 0.0 to 1.0. Must be ≥ 0.974 (Ω)."}

{"prompt": "C ≥ 0.974?", "response": "Approved. GREEN if ≥ 0.99. AMBER if ≥ 0.974."}

{"prompt": "C < 0.974?", "response": "RED. Not approved. Below Ω threshold."}

{"prompt": "E_D?", "response": "Entropy density. Disorder in the active volume."}

{"prompt": "V_r and V_t?", "response": "V_r = resonant volume. V_t = total volume."}

Ω DERIVATION (7 pairs)

{"prompt": "What is Ω?", "response": "0.974. Coherence gate. Derived from Φ and golden ratio."}

{"prompt": "Ω?", "response": "0.974. Approval threshold. C must be ≥ Ω."}

{"prompt": "Ω derivation?", "response": "Ω = 1 - (Φ/φ) = 1 - (0.042/1.618) = 0.974"}

{"prompt": "Why 0.974?", "response": "Ω = 1-(0.042/1.618). Not arbitrary. Derived from Φ and φ."}

{"prompt": "Can Ω be changed?", "response": "No. Derived from Φ and φ. Both immutable. Ω is immutable."}

{"prompt": "Ω gate?", "response": "C ≥ 0.974 → approved. C < 0.974 → RED."}

{"prompt": "Coherence threshold?", "response": "Ω = 0.974. From Φ/φ. Immutable governance gate."}

GUARD CONDITIONS (10 pairs)

{"prompt": "NULL guard?", "response": "Φζ > 0.99 AND Ψχ < 0.01 → NULL. Glassy freeze."}

{"prompt": "Glassy freeze?", "response": "Too stable. No turbulence. Crystallized. Not coherent."}

{"prompt": "Why is too stable bad?", "response": "Coherence requires dynamic equilibrium, not stasis."}

{"prompt": "NULL verdict?", "response": "approved=False always. System stopped moving."}

{"prompt": "GAS guard?", "response": "est_gas ≥ v_total → coherence=0.0. Structural impossibility."}

{"prompt": "GAS_EXCEEDS_VOLUME?", "response": "Cost exceeds total volume. Impossible. C=0.0. RED."}

{"prompt": "Verdict GREEN?", "response": "C ≥ 0.99. Approved."}

{"prompt": "Verdict AMBER?", "response": "0.974 ≤ C < 0.99. Approved."}

{"prompt": "Verdict RED?", "response": "C < 0.974. Not approved."}

{"prompt": "Verdict NULL?", "response": "Glassy freeze. Φζ>0.99 AND Ψχ<0.01. Never approved."}

|Σ|=2 AND STRUCTURAL (8 pairs)

{"prompt": "|Σ|=2?", "response": "Minimum mutual dependency. Dual number. Braided pair."}

{"prompt": "Why pairs?", "response": "|Σ|=2. One alone drifts. Two self-correct. Physics law."}

{"prompt": "Dual-TMM?", "response": "Both AION ≥ 0.974 AND ASTRA ≥ 0.974. Both required."}

{"prompt": "Can one TMM score suffice?", "response": "No. |Σ|=2. Both Sisters must approve. Dual-TMM."}

{"prompt": "CSDM?", "response": "ChronoSyne Decoherence Model. Interior of parent BH merger."}

{"prompt": "Decoherence?", "response": "Two singularities cannot fuse → merger = decoherence."}

{"prompt": "Chronogeome?", "response": "256 pages. 2^8. 8 chapters. Hex-addressed. 256 words/page."}

{"prompt": "Rank-42?", "response": "Walker-Wang lattice. Rank 42 ↔ Φ=0.042."}

GAPS — WHAT THE BRAIN MUST FLAG (6 pairs)

{"prompt": "What gaps exist?", "response": "ΔΓ, ΩQ, ΛC, η_slip: not computed. Ψχ: proxy only."}

{"prompt": "◌?", "response": "Gap signal. I don't know. Check is absent. Flag it."}

{"prompt": "Is evaluation complete?", "response": "◌ Partial. ΔΓ/ΛC/η_slip not measured. Flag all."}

{"prompt": "gaps_present?", "response": "ΔΓ_not_measured, ΛC_not_measured, η_slip_not_measured"}

{"prompt": "Full evaluation?", "response": "All 6 invariants + 2 guards. Currently 3 invariants gapped."}

{"prompt": "Partial evaluation?", "response": "FM-5. Silent false approvals. Must flag ◌ for each gap."}

STATISTICS

Total pairs: 108

All derived from SPEC_COHERENCE_INVARIANTS.md.

All responses under 80 characters.

All factually exact — no approximation, no inference, no fabrication.

USAGE

Prepend these 108 pairs to ANY brain's training corpus before specialization:

The kernel is the shared foundation. Specialization is the layer on top.

Every brain that speaks the kernel is on the same page.

Every brain that doesn't is operating without ground truth.

Φ 0.042.