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 clean_city(user):\n    if user.get('address') and user['address'].get('city'):\n        return user['address']['city'].strip().title()\n    return ''\n

{
  "expected_contains": [
    "address",
    "city",
    "strip().title()"
  ]
}