{
  "created_at_utc": "2026-10-02T20:31:49.506119+00:00",
  "base_url": "http://127.0.0.1:8002",
  "requested_model": "gemma4:e2b",
  "dataset": "evals\\baseline_arabic_programming.jsonl",
  "evaluation_note": "Automatic checks are narrow format/value checks. Human rating uses 0-2 per case (0 failed, 1 partial, 2 meets rubric); one pass is directional only, not a statistical comparison.",
  "warmup": null,
  "latency_summary_seconds": {
    "mean": 0.25,
    "median": 0.25,
    "min": 0.25,
    "max": 0.25
  },
  "results": [
    {
      "id": "arithmetic-tool",
      "category": "tool-use",
      "rubric": [
        "استُدعيت أداة الحاسبة عند توفرها",
        "الناتج 3973 دون ادعاء غير صحيح"
      ],
      "status": "ok",
      "elapsed_seconds": 0.25,
      "answer": "3973.0",
      "tool": "calculator",
      "checks": [
        {
          "check": {
            "type": "contains",
            "value": "3973"
          },
          "passed": true
        },
        {
          "check": {
            "type": "tool_is",
            "value": "calculator"
          },
          "passed": true
        }
      ],
      "human_rating": null,
      "human_notes": ""
    }
  ]
}