{
  "evidence": {
    "runId": "e243b03f-ce07-4829-a164-68d859dbe85c",
    "evaluatedAt": "2026-09-27T21:21:33.447Z",
    "packageVersion": "0.8.1",
    "recipeFingerprint": "4740e3077ceb0727eb06a35d4c4e3c56e01da0c117fbc378b8f7e940b53c9bfa",
    "datasetFingerprint": "c5a38a7a68e9e4b4ff2661beccfb6aaee58f5b087f35b25254e7a1cc9b12ec26",
    "split": "held-out",
    "mode": "live",
    "policy": {
      "minConfidence": 0.8,
      "selectedOn": {
        "runId": "0f417565-224c-4e07-86d5-f80e27bffc8e",
        "datasetFingerprint": "e1a6928c7c05552d47a930818c0b81130f2dad2ca64407cdb5c4d88a318bf4b3",
        "recipeFingerprint": "4740e3077ceb0727eb06a35d4c4e3c56e01da0c117fbc378b8f7e940b53c9bfa",
        "model": "typesafe-ai/jev",
        "split": "development"
      }
    },
    "provenance": [
      {
        "method": "public-dataset",
        "source": "https://raw.githubusercontent.com/nodejs/node/v22.20.0/doc/api/async_context.md; source Markdown labels, transformed by build-ingestion-cases.mjs; no independent human review; Node.js MIT license.",
        "cases": 8
      },
      {
        "method": "public-dataset",
        "source": "https://raw.githubusercontent.com/nodejs/node/v22.20.0/doc/api/buffer.md; source Markdown labels, transformed by build-ingestion-cases.mjs; no independent human review; Node.js MIT license.",
        "cases": 8
      },
      {
        "method": "public-dataset",
        "source": "https://raw.githubusercontent.com/nodejs/node/v22.20.0/doc/api/cli.md; source Markdown labels, transformed by build-ingestion-cases.mjs; no independent human review; Node.js MIT license.",
        "cases": 8
      },
      {
        "method": "public-dataset",
        "source": "https://raw.githubusercontent.com/nodejs/node/v22.20.0/doc/api/console.md; source Markdown labels, transformed by build-ingestion-cases.mjs; no independent human review; Node.js MIT license.",
        "cases": 8
      },
      {
        "method": "public-dataset",
        "source": "https://raw.githubusercontent.com/nodejs/node/v22.20.0/doc/api/events.md; source Markdown labels, transformed by build-ingestion-cases.mjs; no independent human review; Node.js MIT license.",
        "cases": 8
      },
      {
        "method": "public-dataset",
        "source": "https://raw.githubusercontent.com/nodejs/node/v22.20.0/doc/api/fs.md; source Markdown labels, transformed by build-ingestion-cases.mjs; no independent human review; Node.js MIT license.",
        "cases": 8
      },
      {
        "method": "public-dataset",
        "source": "https://raw.githubusercontent.com/nodejs/node/v22.20.0/doc/api/http.md; source Markdown labels, transformed by build-ingestion-cases.mjs; no independent human review; Node.js MIT license.",
        "cases": 8
      },
      {
        "method": "public-dataset",
        "source": "https://raw.githubusercontent.com/nodejs/node/v22.20.0/doc/api/https.md; source Markdown labels, transformed by build-ingestion-cases.mjs; no independent human review; Node.js MIT license.",
        "cases": 8
      },
      {
        "method": "public-dataset",
        "source": "https://raw.githubusercontent.com/nodejs/node/v22.20.0/doc/api/net.md; source Markdown labels, transformed by build-ingestion-cases.mjs; no independent human review; Node.js MIT license.",
        "cases": 8
      },
      {
        "method": "public-dataset",
        "source": "https://raw.githubusercontent.com/nodejs/node/v22.20.0/doc/api/path.md; source Markdown labels, transformed by build-ingestion-cases.mjs; no independent human review; Node.js MIT license.",
        "cases": 8
      },
      {
        "method": "public-dataset",
        "source": "https://raw.githubusercontent.com/nodejs/node/v22.20.0/doc/api/perf_hooks.md; source Markdown labels, transformed by build-ingestion-cases.mjs; no independent human review; Node.js MIT license.",
        "cases": 8
      },
      {
        "method": "public-dataset",
        "source": "https://raw.githubusercontent.com/nodejs/node/v22.20.0/doc/api/querystring.md; source Markdown labels, transformed by build-ingestion-cases.mjs; no independent human review; Node.js MIT license.",
        "cases": 8
      },
      {
        "method": "public-dataset",
        "source": "https://raw.githubusercontent.com/nodejs/node/v22.20.0/doc/api/readline.md; source Markdown labels, transformed by build-ingestion-cases.mjs; no independent human review; Node.js MIT license.",
        "cases": 8
      },
      {
        "method": "public-dataset",
        "source": "https://raw.githubusercontent.com/nodejs/node/v22.20.0/doc/api/stream.md; source Markdown labels, transformed by build-ingestion-cases.mjs; no independent human review; Node.js MIT license.",
        "cases": 8
      },
      {
        "method": "public-dataset",
        "source": "https://raw.githubusercontent.com/nodejs/node/v22.20.0/doc/api/string_decoder.md; source Markdown labels, transformed by build-ingestion-cases.mjs; no independent human review; Node.js MIT license.",
        "cases": 7
      },
      {
        "method": "public-dataset",
        "source": "https://raw.githubusercontent.com/nodejs/node/v22.20.0/doc/api/test.md; source Markdown labels, transformed by build-ingestion-cases.mjs; no independent human review; Node.js MIT license.",
        "cases": 8
      },
      {
        "method": "public-dataset",
        "source": "https://raw.githubusercontent.com/nodejs/node/v22.20.0/doc/api/tty.md; source Markdown labels, transformed by build-ingestion-cases.mjs; no independent human review; Node.js MIT license.",
        "cases": 8
      },
      {
        "method": "public-dataset",
        "source": "https://raw.githubusercontent.com/nodejs/node/v22.20.0/doc/api/url.md; source Markdown labels, transformed by build-ingestion-cases.mjs; no independent human review; Node.js MIT license.",
        "cases": 8
      },
      {
        "method": "public-dataset",
        "source": "https://raw.githubusercontent.com/nodejs/node/v22.20.0/doc/api/v8.md; source Markdown labels, transformed by build-ingestion-cases.mjs; no independent human review; Node.js MIT license.",
        "cases": 8
      },
      {
        "method": "public-dataset",
        "source": "https://raw.githubusercontent.com/nodejs/node/v22.20.0/doc/api/worker_threads.md; source Markdown labels, transformed by build-ingestion-cases.mjs; no independent human review; Node.js MIT license.",
        "cases": 8
      },
      {
        "method": "public-dataset",
        "source": "https://raw.githubusercontent.com/nodejs/node/v22.20.0/doc/api/zlib.md; source Markdown labels, transformed by build-ingestion-cases.mjs; no independent human review; Node.js MIT license.",
        "cases": 8
      }
    ]
  },
  "recipe": "paragraph-boundary",
  "model": "typesafe-ai/jev",
  "models": [
    "typesafe-ai/jev"
  ],
  "cases": 167,
  "completed": 167,
  "failed": 0,
  "correct": 155,
  "wrong": 12,
  "ready": 82,
  "review": 85,
  "accuracy": 0.928,
  "completedAccuracy": 0.928,
  "accuracyInterval95": [
    0.8786,
    0.9584
  ],
  "readyAccuracy": 0.988,
  "readyAccuracyInterval95": [
    0.9341,
    0.9978
  ],
  "reviewRate": 0.509,
  "contestedAccuracy": null,
  "adversarialAccuracy": null,
  "calibration": [
    {
      "confidence": "0-0.5",
      "cases": 47,
      "accuracy": 0.766
    },
    {
      "confidence": "0.5-0.6",
      "cases": 14,
      "accuracy": 1
    },
    {
      "confidence": "0.6-0.7",
      "cases": 11,
      "accuracy": 1
    },
    {
      "confidence": "0.7-0.8",
      "cases": 13,
      "accuracy": 1
    },
    {
      "confidence": "0.8-0.9",
      "cases": 6,
      "accuracy": 0.833
    },
    {
      "confidence": "0.9-1",
      "cases": 76,
      "accuracy": 1
    }
  ],
  "thresholdEvaluation": "frozen-policy",
  "thresholds": [],
  "suggestedMinConfidence": null,
  "latencyMs": {
    "p50": 242.82,
    "p95": 347.69,
    "totalCaseTime": 42641.84
  },
  "usage": {
    "requests": 167,
    "input_tokens": 90864,
    "output_tokens": 6993,
    "unknownResponses": 0,
    "failedRequests": 0,
    "models": [
      "typesafe-ai/jev"
    ]
  },
  "cost": {
    "model": "typesafe-ai/jev",
    "inputPerMillion": 0.042,
    "outputPerMillion": 0,
    "currency": "USD",
    "date": "2026-09-27",
    "source": "https://ai-gateway.vercel.sh/v1/models",
    "estimated": 0.00381629
  },
  "confusion": {
    "{\"verdict\":\"continue\"}": {
      "total": 84,
      "correct": 83,
      "misses": {
        "{\"verdict\":\"separate\"}": 1
      }
    },
    "{\"verdict\":\"separate\"}": {
      "total": 83,
      "correct": 72,
      "misses": {
        "{\"verdict\":\"continue\"}": 11
      }
    }
  },
  "failures": [
    {
      "id": "node-v22.20.0-async_context.md-separate-128",
      "expected": {
        "verdict": "separate"
      },
      "actual": {
        "verdict": "continue"
      },
      "confidence": 0.42
    },
    {
      "id": "node-v22.20.0-console.md-separate-71",
      "expected": {
        "verdict": "separate"
      },
      "actual": {
        "verdict": "continue"
      },
      "confidence": 0.25
    },
    {
      "id": "node-v22.20.0-http.md-separate-208",
      "expected": {
        "verdict": "separate"
      },
      "actual": {
        "verdict": "continue"
      },
      "confidence": 0.34
    },
    {
      "id": "node-v22.20.0-http.md-separate-230",
      "expected": {
        "verdict": "separate"
      },
      "actual": {
        "verdict": "continue"
      },
      "confidence": 0.38
    },
    {
      "id": "node-v22.20.0-net.md-continue-52",
      "expected": {
        "verdict": "continue"
      },
      "actual": {
        "verdict": "separate"
      },
      "confidence": 0.88
    },
    {
      "id": "node-v22.20.0-readline.md-separate-68",
      "expected": {
        "verdict": "separate"
      },
      "actual": {
        "verdict": "continue"
      },
      "confidence": 0.37
    },
    {
      "id": "node-v22.20.0-url.md-separate-80",
      "expected": {
        "verdict": "separate"
      },
      "actual": {
        "verdict": "continue"
      },
      "confidence": 0.34
    },
    {
      "id": "node-v22.20.0-v8.md-separate-34",
      "expected": {
        "verdict": "separate"
      },
      "actual": {
        "verdict": "continue"
      },
      "confidence": 0.45
    },
    {
      "id": "node-v22.20.0-v8.md-separate-92",
      "expected": {
        "verdict": "separate"
      },
      "actual": {
        "verdict": "continue"
      },
      "confidence": 0.22
    },
    {
      "id": "node-v22.20.0-zlib.md-separate-220",
      "expected": {
        "verdict": "separate"
      },
      "actual": {
        "verdict": "continue"
      },
      "confidence": 0.41
    },
    {
      "id": "node-v22.20.0-zlib.md-separate-486",
      "expected": {
        "verdict": "separate"
      },
      "actual": {
        "verdict": "continue"
      },
      "confidence": 0.3
    },
    {
      "id": "node-v22.20.0-zlib.md-separate-496",
      "expected": {
        "verdict": "separate"
      },
      "actual": {
        "verdict": "continue"
      },
      "confidence": 0.3
    }
  ],
  "documentEvaluation": {
    "baseline": "Frozen syntax/punctuation rules in evals/lib/ingestion-metrics.mjs",
    "baselineAccuracy": 0.9760479041916168,
    "documentCount": 21,
    "perLabel": [
      {
        "label": "continue",
        "cases": 84,
        "precision": 0.8829787234042553,
        "recall": 0.9880952380952381,
        "f1": 0.9325842696629213
      },
      {
        "label": "separate",
        "cases": 83,
        "precision": 0.9863013698630136,
        "recall": 0.8674698795180723,
        "f1": 0.9230769230769231
      }
    ],
    "macroF1": 0.9278305963699223,
    "readyCoverage": 0.49101796407185627,
    "acceptedDecisions": 81,
    "acceptedDocuments": 21,
    "acceptedPrecision": 1,
    "acceptedPrecisionInterval95": [
      1,
      1
    ],
    "acceptedContinuationRecall": 0.9642857142857143,
    "baselineImprovementInterval95": [
      -0.10119047619047619,
      0
    ],
    "intervalMethod": "Deterministic percentile bootstrap of whole source documents, 2000 replicates. Sparse/all-correct samples can yield degenerate intervals; these are descriptive, not guarantees."
  },
  "acceptance": {
    "met": false,
    "checks": {
      "noProviderFailures": true,
      "beatsFrozenBaseline": false,
      "acceptedJoinPrecision": true,
      "precisionLowerBound": true,
      "continuationRecall": true,
      "reviewRate": false,
      "acceptedJoins": false,
      "acceptedDocuments": false
    },
    "label": "Experimental: declared document acceptance policy not met"
  },
  "gatewayIdentity": "Gateway returned the alias typesafe-ai/jev; backend version and BYOK credential use are not independently established."
}
