A0 integral-control checks
Synthetic, dimensionless models; no organism result or H* verdict.
Gates: 24/24 passed.
| Measurement | Result |
|---|---|
| A0 circuit: endpoint vs direct development (max abs, log units) | 2.1e-10 |
| A0 circuit: endpoints vs linear map P dz0, same samples (relative RMS) | 0.0293 |
| A0 circuit: covariance vs P S0 P^T (relative) / 95% sampling band | 0.122 / 0.164 |
| A0 circuit: variance of sensor-controlled combinations / initial | 0.000519 |
| Two-sensor neuron: endpoints vs linear map P dz0 (relative RMS) | 0.000703 |
| Two-sensor neuron: covariance error, no jitter / with jitter / 95% sampling band | 0.108 / 0.105 / 0.174 |
| Two-sensor neuron: J Cov J^T vs set-point jitter covariance (relative) | 0.0507 |
| Unstable root: sensor residual / loop-gain eigenvalues | 5.55e-17 / 0.74, -0.732 |
| Unstable root: distance of developed endpoint from it | 122 |
| History, constant gain: adult difference after the transient manipulation | 8.66e-15 |
| History, non-integrable gain: adult difference, excitation / mild silencing (involutivity defect) | 0.0239 / 0.00011 (0.0329) |
| Gauge G -> G C: endpoint difference / recovery half-times | 6.37e-11 / 0.75 vs 0.2 |
| Implicit gradient vs developed-endpoint finite differences (max abs) | 1.32e-06 |
| Gate | Passed |
|---|---|
| A0 endpoints match direct development | yes |
| A0 endpoints follow the linear map P dz0 (<5%) | yes |
| A0 covariance within sampling band of P S0 P^T | yes |
| A0 controlled combinations suppressed (<1%) | yes |
| two-sensor slice solve matches development | yes |
| two-sensor endpoints follow the linear map P dz0 (<5%) | yes |
| two-sensor covariance within sampling band | yes |
| covariance with set-point jitter within 2x sampling band | yes |
| J Cov J^T equals jitter covariance (<2x sampling band) | yes |
| unstable root solves the sensor equations | yes |
| unstable root flagged by loop gain | yes |
| development does not reach the unstable root | yes |
| continuation does not report the unstable root | yes |
| constant gain conserves N^T z | yes |
| constant gain is involutive | yes |
| constant gain: no lasting history effect | yes |
| non-integrable gain detected | yes |
| non-integrable gain: lasting history effect | yes |
| both reach the set points | yes |
| gauge: G C leaves endpoint unchanged | yes |
| gauge: G C changes recovery time | yes |
| implicit gradients match development FD | yes |
| augmented det identity | yes |
| augmented singular when JG singular | yes |
Seed 11. Full values: report.json.