def alpaca_evaluator()

in ml/eval/alpaca.py [0:0]

• 7 lines of code
• 2 McCabe index (conditional complexity)

def alpaca_evaluator(model_name, num_samples=200):
    results = run_evaluation(
        model=model_name,
        num_samples=num_samples,  # fewer samples for quick testing
        reference_model="gpt-4",  # Compare against GPT-4 (optional)
    )
    return results