{
  "status": "complete",
  "benchmark": "humaneval",
  "anchor": {
    "model": "Qwen/Qwen2.5-Coder-7B",
    "published": {
      "score": 61.6,
      "metric": "pass@1",
      "source": "Qwen2.5-Coder Technical Report Table 5, arXiv:2409.12186"
    },
    "measured_scores": {
      "humaneval": 62.2,
      "humaneval_plus": 53.7
    },
    "delta": 0.6
  },
  "model_under_test": {
    "path": "/home/joel/forge_v2_qwen7b_compensated_kl",
    "measured_scores": {
      "humaneval": 61.0,
      "humaneval_plus": 53.0
    }
  },
  "tolerance": 3.0,
  "calibration_passed": true
}