sft-qwen2.5-1.5b/eval-results/mbpp/metrics.json

{
  "mbpp": {
    "pass@1": {
      "num_entries": 378,
      "avg_tokens": 66,
      "gen_seconds": 20,
      "passing_base_tests": 60.58201058201058,
      "passing_plus_tests": 51.851851851851855
    }
  }
}