> ## Documentation Index
> Fetch the complete documentation index at: https://docs.thrindex.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Conformance Envelope V0

> The ratified CONFORMANCE_ENVELOPE_V0 constants, their derivation, scope, and reopening conditions.

## The constants

```python theme={null}
CONFORMANCE_ENVELOPE_V0 = {
    "status": "Final",        # ratified 2026-07-10
    "version": "v0",
    "T_mean_threshold":   0.020,
    "T_max_threshold":    0.130,
    "pred_agreement_min": 0.900,
    "min_test_samples":   100,
}
```

A backend must satisfy **all three simultaneously** to earn the THRINDEX Certified \[v0] badge:

| Metric        | Threshold | Meaning                                 |
| ------------- | --------- | --------------------------------------- |
| `T_mean`      | ≤ 0.020   | Mean spike-count rate error ≤ 2%        |
| `T_max` (p99) | ≤ 0.130   | Worst-neuron error ≤ 13%                |
| `P_pred`      | ≥ 0.900   | ≥ 90% of classification decisions agree |

## Fixture set

The ratification fixture set is frozen at CRC32 `e2ebd845` (120 SHD test samples, stratified 6 per class × 20 classes). It lives at `conformance/fixtures/shd_ratify_v1/frozen/` in the monorepo.

| Property            | Value                                     |
| ------------------- | ----------------------------------------- |
| Fixture fingerprint | CRC32 `e2ebd845`, 120 files               |
| Selection           | Stratified: 6 samples × 20 SHD classes    |
| Source              | SHD test split (CC BY 4.0)                |
| Generation script   | `scripts/freeze_shd_fixtures.py --seed 0` |

## Quantization model

The envelope was calibrated against per-channel symmetric int8 — the hardware-class quantization that real neuromorphic digital chips implement:

* Scale per output channel: `scale_n = max_abs(row_n) / 127`
* Rounding: round-half-to-even (banker's rounding, matches `thrindex-numerics`)
* Clamp: \[-127, 127]; dequantize: `q as f32 × scale_n`
* Biases: left in f32

## Ratification evidence

Measured on the frozen 120-sample fixture set. Reference model: `templates/keyword-spotting/model.thx` (Dense 700→512→20, τ\_mem=20ms, 64.66% SHD test accuracy).

**INT8 per-channel — should PASS (this is the hardware model the envelope targets):**

| Metric        | Observed  | V0 threshold | Headroom |
| ------------- | --------- | ------------ | -------- |
| `T_mean`      | 5.28×10⁻³ | ≤ 0.020      | **3.8×** |
| `T_max` (p99) | 1.00×10⁻¹ | ≤ 0.130      | **30%**  |
| `P_pred`      | 0.983     | ≥ 0.900      | **8.3%** |

**INT4 per-channel — should FAIL (confirms thresholds are not vacuously loose):**

| Metric        | Observed  | V0 threshold | int4/int8 ratio | Verdict                |
| ------------- | --------- | ------------ | --------------- | ---------------------- |
| `T_mean`      | 8.38×10⁻² | ≤ 0.020      | 15.9×           | **FAILS** (4.2× above) |
| `T_max` (p99) | 9.20×10⁻¹ | ≤ 0.130      | 9.2×            | **FAILS** (7.1× above) |
| `P_pred`      | 0.617     | ≥ 0.900      | —               | **FAILS** (46% below)  |

The separation is decisive. There is no metric where int8 and int4 are ambiguous.

## Threshold derivation

**T\_mean = 0.020** — the provisional value was confirmed by data. INT8 sits at 26% of the limit (3.8× headroom); INT4 at 419% (4.2× above). The 15.9× int8/int4 gap means the threshold is well-placed: tight enough to catch INT4 decisively, loose enough to accommodate INT8 with margin.

**T\_max = 0.130** — the provisional value (0.100) equaled the observed INT8 `agg_max` exactly (zero margin). One extreme sample had a neuron differing by 10 spikes over 100 timesteps (rate error = 0.10). Adding 30% headroom gives 0.130. The auto-derived 40%-into-gap value (0.428) was rejected as too loose — certifying a backend with worst-neuron divergence of 43% has no meaningful inference interpretation. 0.130 passes INT8 with margin and fails INT4 (0.920) by 7.1×.

**P\_min = 0.900** — the provisional value (0.990) was unreachable: INT8's observed overall agreement was 0.983, with boundary-case agreement 0.9375 (15/16 near-tie samples). 0.900 seats below both with 8.3% and 4.2% headroom respectively. INT4 at 0.617 fails by 32%. Semantic interpretation: the backend must agree with the float reference on ≥90% of test inputs.

## V0 scope

**This envelope was ratified against one reference model** (SHD keyword spotting, Dense 700→512→20, 64.66% accuracy, 120 frozen test samples). It brackets per-channel int8 divergence on that specific architecture and weight distribution. It is not yet a universal constant.

A backend certified under V0 carries the badge **THRINDEX Certified \[v0]**.

## Reopening triggers

Any one of the following mandates a V1 measurement:

1. A structurally different model enters the conformance suite (recurrent layers, different T, Conv2d-heavy, or substantially different weight distribution)
2. A real silicon backend is added whose quantization characteristics differ from per-channel int8
3. A latency-code or time-to-first-spike model enters the suite

## Re-certification

Tightening any threshold is a breaking change: V0-certified backends must be re-certified under V1. Loosening requires the same approval path. The harness supports `--envelope v0` indefinitely for retesting. The V0 badge is retained in published history but marked superseded when V1 ships.