Allows submission of a “golden set” of question-answer pairs to evaluate the accuracy and relevance of one or more knowledge bases. The request body must contain two parallel arrays: questions and goldenAnswers, where element N in questions corresponds to element N in goldenAnswers. The service will execute each question against the specified knowledge base(s), compute matching scores against the provided golden answers, and return detailed evaluation results including the model’s top answer and a numeric score.