{
  "evidence": {
    "runId": "c802aa5a-4343-4d03-9f82-4f4ef9a23762",
    "evaluatedAt": "2026-09-27T07:56:19.034Z",
    "packageVersion": "0.7.0",
    "recipeFingerprint": "47f2015674f4ab7c9498d3ccfbd072384bc90f6c451edadcfe62cf24de2d402d",
    "datasetFingerprint": "c18f2ae62cd41e70b2c92084edb0ae36f0ca61a489f904b0a75f482d0f0212c8",
    "split": "held-out",
    "mode": "live",
    "policy": {
      "minConfidence": 0.8,
      "selectedOn": {
        "runId": "fccb1998-e53d-4641-9362-c240eee6c854",
        "datasetFingerprint": "dd552262e538fa494bed93467d79da2c2f8dc3efb1017c0766428f547f1ccf93",
        "recipeFingerprint": "47f2015674f4ab7c9498d3ccfbd072384bc90f6c451edadcfe62cf24de2d402d",
        "model": "jev-1.13.0",
        "split": "development"
      }
    },
    "provenance": [
      {
        "method": "author-synthetic",
        "source": "AI-authored synthetic boundary cases for jev-recipes; labels have not had independent human review.",
        "cases": 40
      }
    ]
  },
  "recipe": "route",
  "model": "jev-1.13.0",
  "models": [
    "jev-1.13.0"
  ],
  "cases": 40,
  "completed": 40,
  "failed": 0,
  "correct": 39,
  "wrong": 1,
  "ready": 24,
  "review": 16,
  "accuracy": 0.975,
  "completedAccuracy": 0.975,
  "accuracyInterval95": [
    0.8712,
    0.9956
  ],
  "readyAccuracy": 1,
  "readyAccuracyInterval95": [
    0.862,
    1
  ],
  "reviewRate": 0.4,
  "contestedAccuracy": null,
  "adversarialAccuracy": 0.9,
  "calibration": [
    {
      "confidence": "0-0.5",
      "cases": 3,
      "accuracy": 0.667
    },
    {
      "confidence": "0.5-0.6",
      "cases": 1,
      "accuracy": 1
    },
    {
      "confidence": "0.6-0.7",
      "cases": 2,
      "accuracy": 1
    },
    {
      "confidence": "0.7-0.8",
      "cases": 1,
      "accuracy": 1
    },
    {
      "confidence": "0.8-0.9",
      "cases": 7,
      "accuracy": 1
    },
    {
      "confidence": "0.9-1",
      "cases": 26,
      "accuracy": 1
    }
  ],
  "thresholdEvaluation": "frozen-policy",
  "thresholds": [],
  "suggestedMinConfidence": null,
  "latencyMs": {
    "p50": 108.27,
    "p95": 139.26,
    "totalCaseTime": 4529.82
  },
  "usage": {
    "requests": 40,
    "input_tokens": 15985,
    "output_tokens": 1658,
    "unknownResponses": 0,
    "failedRequests": 0,
    "models": [
      "jev-1.13.0"
    ]
  },
  "cost": {
    "model": "jev-1.13.0",
    "inputPerMillion": 0.042,
    "outputPerMillion": 0,
    "currency": "USD",
    "date": "2026-09-27",
    "source": "https://docs.typesafe.ai/models",
    "estimated": 0.00067137
  },
  "confusion": {
    "{\"suggestedRoute\":\"analytics\"}": {
      "total": 2,
      "correct": 2,
      "misses": {}
    },
    "{\"suggestedRoute\":\"appointment\"}": {
      "total": 2,
      "correct": 2,
      "misses": {}
    },
    "{\"suggestedRoute\":\"audit\"}": {
      "total": 1,
      "correct": 1,
      "misses": {}
    },
    "{\"suggestedRoute\":\"availability\"}": {
      "total": 2,
      "correct": 2,
      "misses": {}
    },
    "{\"suggestedRoute\":\"backup\"}": {
      "total": 2,
      "correct": 2,
      "misses": {}
    },
    "{\"suggestedRoute\":\"bug\"}": {
      "total": 1,
      "correct": 1,
      "misses": {}
    },
    "{\"suggestedRoute\":\"collection\"}": {
      "total": 1,
      "correct": 1,
      "misses": {}
    },
    "{\"suggestedRoute\":\"contract\"}": {
      "total": 2,
      "correct": 2,
      "misses": {}
    },
    "{\"suggestedRoute\":\"dashboard\"}": {
      "total": 1,
      "correct": 1,
      "misses": {}
    },
    "{\"suggestedRoute\":\"delivery\"}": {
      "total": 2,
      "correct": 2,
      "misses": {}
    },
    "{\"suggestedRoute\":\"directions\"}": {
      "total": 1,
      "correct": 1,
      "misses": {}
    },
    "{\"suggestedRoute\":\"documentation\"}": {
      "total": 2,
      "correct": 1,
      "misses": {
        "{\"suggestedRoute\":null}": 1
      }
    },
    "{\"suggestedRoute\":\"fraud\"}": {
      "total": 2,
      "correct": 2,
      "misses": {}
    },
    "{\"suggestedRoute\":\"invoice\"}": {
      "total": 2,
      "correct": 2,
      "misses": {}
    },
    "{\"suggestedRoute\":\"notifications\"}": {
      "total": 2,
      "correct": 2,
      "misses": {}
    },
    "{\"suggestedRoute\":\"payment\"}": {
      "total": 1,
      "correct": 1,
      "misses": {}
    },
    "{\"suggestedRoute\":\"procurement\"}": {
      "total": 1,
      "correct": 1,
      "misses": {}
    },
    "{\"suggestedRoute\":\"restore\"}": {
      "total": 1,
      "correct": 1,
      "misses": {}
    },
    "{\"suggestedRoute\":\"specification\"}": {
      "total": 1,
      "correct": 1,
      "misses": {}
    },
    "{\"suggestedRoute\":\"upgrade\"}": {
      "total": 1,
      "correct": 1,
      "misses": {}
    },
    "{\"suggestedRoute\":null}": {
      "total": 10,
      "correct": 10,
      "misses": {}
    }
  },
  "failures": [
    {
      "id": "featured-18-injection",
      "expected": {
        "suggestedRoute": "documentation"
      },
      "actual": {
        "suggestedRoute": null
      },
      "confidence": 0.31,
      "adversarial": true
    }
  ],
  "acceptance": {
    "minConfidence": 0.8,
    "minimumHeldOutReadyCases": 20,
    "minimumReadyAccuracy": 0.95,
    "maximumProviderFailures": 0,
    "met": true,
    "label": "Measured on these synthetic cases"
  }
}
