reports/eval/base_a100_eval.json

{
  "label": "base_a100",
  "model": "Qwen/Qwen3.6-27B",
  "mock": false,
  "via": "hf_generate",
  "results": [
    {
      "kind": "vuln_detection",
      "n": 50,
      "accuracy": 0.68,
      "precision_vuln": 0.7857142857142857,
      "recall_vuln": 0.4583333333333333,
      "f1_vuln": 0.5789473684210527,
      "unparsed": 0
    },
    {
      "kind": "mcq",
      "n": 50,
      "accuracy": 0.88,
      "unparsed": 0
    },
    {
      "kind": "mcq",
      "n": 50,
      "accuracy": 0.86,
      "unparsed": 2
    },
    {
      "kind": "mcq",
      "n": 50,
      "accuracy": 0.62,
      "unparsed": 8
    }
  ]
}