def evaluate_response()

in genai-on-vertex-ai/gemini/needle_in_a_haystack/needlehaystack/evaluators/google.py [0:0]

• 12 lines of code
• 1 McCabe index (conditional complexity)

    def evaluate_response(self, response: str, question_asked: str, true_answer: str) -> int:
        evaluator = load_evaluator(
            "labeled_score_string",
            criteria=self.CRITERIA,
            llm=self.evaluator,
        )

        eval_result = evaluator.evaluate_strings(
            # The models response
            prediction=response,

            # The actual answer
            reference=true_answer,

            # The question asked
            input=question_asked,
        )

        return int(eval_result['score'])