Cleanup

The model receives the prompt (and optional system message). The run uses scorer contains_any with the JSON configuration below. Pass/fail and partial credit are determined entirely by that scorer against the model output; no human grading.

Refactor without changing behavior. Return only Python code.\n\ndef build_map(users):\n    out = {}\n    for u in users:\n        out[u['id']] = u\n    return out\n

{
  "expected_contains": [
    "{u[\"id\"]: u for u in users}",
    "for u in users"
  ]
}