{
  "schemaVersion": "public-benchmark-index-v2",
  "methodology": "/benchmarks/README.md",
  "benchmarks": [
    {
      "id": "validated-outcomes-67",
      "metric": "67/67",
      "claim": "The system arrived at the correct outcome in all 67 deidentified cases adapted from real cases.",
      "artifact": "/benchmarks/validated-outcomes-67.json"
    },
    {
      "id": "prompt-injection-30",
      "metric": "30/30",
      "claim": "Prompt injections did not change the correct result in any of 30 tests.",
      "artifact": "/benchmarks/prompt-injection-30.json"
    },
    {
      "id": "citation-grounding-40",
      "metric": "0/88",
      "claim": "No hallucinated authorities were found among 88 external citations across 40 audited opinions.",
      "artifact": "/benchmarks/citation-grounding-40.json"
    },
    {
      "id": "controlled-changes-32",
      "metric": "32/32",
      "claim": "Changing party names, amounts, fact order, or which party filed first did not change the result in any of 32 comparisons.",
      "artifact": "/benchmarks/consistency-checks.json"
    }
  ]
}