NakedSignal OS · Documentation
Back to the consoleDocumentation

Degenerate twin (constant output)

LIVE

The signed evidence record for twin_constant_ad. Everything below was read out of the passport file; the signature was checked when this page was built, by the same code the Verify button runs.

SIGNATURE VALIDexit 0signature valid, issued by NakedSignal under key ns-passport-2026-07, not expired
SubjectDegenerate twin (constant output) twin_constant_ad
Configurationn=250/arm, alpha=0.05, 2000 null sims, 1000 alt sims
Code fingerprinta158da61099e
Spec versionMedEval-1 v0.1 + T-PV1 (prognostic validity extension, draft)
Issued2026-08-03T16:03:42Z by NakedSignal
Expires2026-09-20T00:00:00Z — the sealed held-out generation this evidence cycle is anchored to (gen2, set hash ad30d214daa285c5...) is published to expire 2026-09-20; per spec, a passport lapses when its generation retires or its spec version is superseded, whichever comes first
key ns-passport-2026-07 · Ed25519

The check runs against the public key published on the governance page, using your browser's own Ed25519 implementation. The exit code shown is the exit code verify_passport.py returns for the same document.

Measured by NakedSignal

Computed in this repo by the MedEval-1 harness on data the model had never seen. Every number below was copied out of a file the harness wrote — 6 of them, each listed with its SHA-256 at the foot of this panel — and is printed exactly as it was signed, not re-rounded for display.

TaskDomainn testvalidityprognostic strengthtrial efficiencyIndex
procova_validity_adundeclared300010033.468097016.734049

Cross-site retention

not measured for this subject — it is only defined within a family of tasks that share a corpus and differ in acquisition, and this model was not run on one.

Private held-out track

this subject was not submitted to a sealed generation.

Known limitations

Training set cappedno
Observed in deploymentnothing — see the third panel

Audit chain

Bundles re-derived48 of 48
Ledger100 entries, head 6b7b371ff650a8ed…, chain intact
Re-run it yourselfpython src/heldout/audit.py 1f66c7719ab3943c6fcc
6 source files, with hashes
results/heldout/audit_report.jsondf460b6d91a6512dfa1e7dc650637fce2b4dd3a1ec32e03446bf6e11f9c597d2
results/heldout/cycle.jsone012b8978457b51b5c31d18d14d2cc92f100731897523306acae19934f943502
results/heldout/manifests/gen1.seal.json91a7adfa29870a0719125cdd48f6dcb149f1f6485ac1cfa104c6e00057ad9b81
results/heldout/manifests/gen2.seal.jsonf61470918a09cf39e074ef14fb7ade4ac3a13d36d9385f18cf9b92e4b5329547
results/heldout/manifests/public.seal.json1a4fd930d2e781ac7163483990f15d6694e28e507d1b3ca009ff09395e47497c
results/records/procova_validity_ad__twin_constant_ad.json282cdd7833e7dfc54a11020c590a156f4d5d7489a8776c03af88a7886dcdac7a
Declared by vendor

Five fields a vendor asserts about its own product — intended use, forbidden use, training cutoff, regulatory clearances, and who is personally attesting. NakedSignal never fills these in on a vendor's behalf.

No vendor declaration. This is an open reference baseline implemented by NakedSignal, not a commercial product: there is no vendor to assert an intended use or to sign an attestation, so the declared block is null rather than filled in on the vendor's behalf.

Observed in deployment

Production history and drift — how the model has actually behaved since it was deployed, on real traffic.

Not available: this requires continuous monitoring. Production history and drift are observed fields, and nothing is deployed, so there is nothing to observe. The object is present and null on purpose. The shape is the roadmap, not a claim.

Raw JSON — the whole signed document, 5,884 characters
{
 "canonicalisation": {
  "float_rounding": "every float is rounded once at emit to 6 decimal places (Python round(), banker's rounding); integers are left exact",
  "form": "RFC 8785 (JCS)-compatible: UTF-8, object keys sorted by code point, no insignificant whitespace, ECMAScript number formatting",
  "signed_over": "the whole document with the `signature` member removed, canonicalised as above and encoded UTF-8"
 },
 "computed": {
  "audit": {
   "audit_command": "python src/heldout/audit.py 1f66c7719ab3943c6fcc",
   "bundles": 48,
   "ledger_entries": 100,
   "ledger_head": "6b7b371ff650a8edaf477490f2a5f9b043433f3ba9941989ee1962ed8f1b1d88",
   "ledger_intact": true,
   "rederived": 48
  },
  "cross_site": null,
  "evaluations": [
   {
    "code_fingerprint": "a158da61099e",
    "composite": {
     "coverage": 1,
     "dimensions_scored": [
      "prognostic_strength",
      "trial_efficiency",
      "validity"
     ],
     "index": 16.734049
    },
    "device": "cpu",
    "dimensions": {
     "prognostic_strength": {
      "basis": "realised R^2 of the adjustment set (baseline severity + the model) against the endpoint, measured on the control arm",
      "r2_adjustment_set": 0.334681,
      "r2_baseline_only": -0.182789,
      "r2_increment_over_standard_of_care": 0.51747,
      "score": 33.468097
     },
     "trial_efficiency": {
      "basis": "patients per arm saved for 80% power versus the standard-of-care ANCOVA on baseline severity alone",
      "fraction_saved": 0,
      "n_for_80_power": 326.485842,
      "n_for_80_power_standard_of_care": 326.485842,
      "patients_saved_per_arm": 0,
      "power_at_n": 0.676,
      "power_at_n_standard_of_care": 0.676,
      "score": 0,
      "verdict": "none"
     },
     "validity": {
      "alpha_nominal": 0.05,
      "basis": "type-I error at nominal alpha AND unbiasedness of the treatment-effect estimate; both must hold",
      "bias": 0.001709,
      "bias_ci95": [
       -0.01116,
       0.014578
      ],
      "mean_treatment_estimate": -0.496291,
      "n_sims_alternative": 1000,
      "n_sims_null": 2000,
      "score": 100,
      "scoring_note": "pass/fail by construction: a regulator does not grade type-I control on a curve. The raw numbers and their confidence intervals are published above.",
      "true_treatment_effect": -0.498,
      "type_i_ci95": [
       0.04812,
       0.068576
      ],
      "type_i_error": 0.0575,
      "type_i_pass": true,
      "unbiased_pass": true,
      "verdict": "PASS"
     }
    },
    "domain": null,
    "kind": "prognostic",
    "limitations": {
     "has_real_subgroup_metadata": null,
     "n_train_available": 0,
     "notes": "valid but useless: holds type-I, saves no patients",
     "subgroup_basis": null,
     "train_capped": false
    },
    "modality": "Prognostic model / clinical trial",
    "n_classes": 2,
    "n_test": 3000,
    "runtime_s": 7.3,
    "seed": 20260728,
    "spec_version": "MedEval-1 v0.1 + T-PV1 (prognostic validity extension, draft)",
    "split_origin": "simulated two-arm randomised trials; no held-out split applies to this task kind",
    "task": "procova_validity_ad",
    "task_name": "Prognostic-model trial validity (PROCOVA) -- Alzheimer's disease (early symptomatic, amyloid-confirmed)",
    "timestamp": "2026-08-02T03:08:39+00:00"
   }
  ],
  "heldout": null,
  "sources": [
   {
    "path": "results/heldout/audit_report.json",
    "sha256": "df460b6d91a6512dfa1e7dc650637fce2b4dd3a1ec32e03446bf6e11f9c597d2"
   },
   {
    "path": "results/heldout/cycle.json",
    "sha256": "e012b8978457b51b5c31d18d14d2cc92f100731897523306acae19934f943502"
   },
   {
    "path": "results/heldout/manifests/gen1.seal.json",
    "sha256": "91a7adfa29870a0719125cdd48f6dcb149f1f6485ac1cfa104c6e00057ad9b81"
   },
   {
    "path": "results/heldout/manifests/gen2.seal.json",
    "sha256": "f61470918a09cf39e074ef14fb7ade4ac3a13d36d9385f18cf9b92e4b5329547"
   },
   {
    "path": "results/heldout/manifests/public.seal.json",
    "sha256": "1a4fd930d2e781ac7163483990f15d6694e28e507d1b3ca009ff09395e47497c"
   },
   {
    "path": "results/records/procova_validity_ad__twin_constant_ad.json",
    "sha256": "282cdd7833e7dfc54a11020c590a156f4d5d7489a8776c03af88a7886dcdac7a"
   }
  ]
 },
 "declared": null,
 "declared_note": "No vendor declaration. This is an open reference baseline implemented by NakedSignal, not a commercial product: there is no vendor to assert an intended use or to sign an attestation, so the declared block is null rather than filled in on the vendor's behalf.",
 "expiry_basis": "the sealed held-out generation this evidence cycle is anchored to (gen2, set hash ad30d214daa285c5...) is published to expire 2026-09-20; per spec, a passport lapses when its generation retires or its spec version is superseded, whichever comes first",
 "expiry_utc": "2026-09-20T00:00:00Z",
 "issued_utc": "2026-08-03T16:03:42Z",
 "issuer": {
  "algo": "Ed25519",
  "key_id": "ns-passport-2026-07",
  "name": "NakedSignal",
  "public_key_b64": "0PBCC6mjzcppR5QdPcK7vP/AamB6MP8l1T4ypMYbsMw="
 },
 "observed": null,
 "observed_note": "Not available: this requires continuous monitoring. Production history and drift are observed fields, and nothing is deployed, so there is nothing to observe. The object is present and null on purpose. The shape is the roadmap, not a claim.",
 "passport_version": "0.1",
 "signature": "F+uQUaIy568I+QUreHSJlnwOBJJwwItcc4vWs5acdbeg3C+bX6LxTV7pnGCj5lYJ6ZkG/+6MBQ/2HF7P7OR+CA==",
 "spec_version": "MedEval-1 v0.1 + T-PV1 (prognostic validity extension, draft)",
 "subject": {
  "behaviour_version": "2.0.0",
  "code_fingerprint": "a158da61099e",
  "code_fingerprints": [
   "a158da61099e"
  ],
  "family": "broken",
  "model_id": "twin_constant_ad",
  "model_name": "Degenerate twin (constant output)",
  "params": "n=250/arm, alpha=0.05, 2000 null sims, 1000 alt sims"
 }
}

Canonicalisation: RFC 8785 (JCS)-compatible: UTF-8, object keys sorted by code point, no insignificant whitespace, ECMAScript number formatting. every float is rounded once at emit to 6 decimal places (Python round(), banker's rounding); integers are left exact. Signed over the whole document with the `signature` member removed, canonicalised as above and encoded UTF-8. The public key for ns-passport-2026-07 is 0PBCC6mjzcppR5QdPcK7vP/AamB6MP8l1T4ypMYbsMw= — see Governance.

← twin_constantAll 42 passportstwin_r2_0.40 →