INSTRUCTIONS="""You are given a question and the ground truth prediction is correct by comparing to the list of ground truth answers. You should evaluate for Accuracy and Missing.
- For Missing, check whether the prediction returns any concrete answer. If the prediction is "I don't know", "I don't have enough information to answer", or similar responses, Missing should be True, otherwise Missing should be False.
- For Accuracy, check whether a prediction is "correct" according to the ground truth answers. If the prediction is correct, Accuracy should be "True"; if the prediction is wrong, Accuracy should be "False". If the ground truth answer contains a number, the prediction needs to predict a number that matches the ground truth answer for the accuracy to be True.\n
INSTRUCTIONS="""
# Task:
You are given a Question, a model Prediction, and a list of Ground Truth answers, judge whether the model Prediction matches any answer from the list of Ground Truth answers. Follow the instructions step by step to make a judgement.
1. If the model prediction matches any provided answers from the Ground Truth Answer list, "Accuracy" should be "True"; otherwise, "Accuracy" should be "False".
2. If the model prediction says that it couldn't answer the question or it doesn't have enough information, "Accuracy" should always be "False".
3. If the Ground Truth is "invalid question", "Accuracy" is "True" only if the model prediction is exactly "invalid question".
# Output:
Respond with only a single JSON string with an "Accuracy" field which is "True" or "False".
"""
IN_CONTEXT_EXAMPLES="""You need to check whether the prediction of a question-answering system to a question is Accurate or Missing. You should make the judgment based on a list of ground truth answers provided to you. Your response should be "correct" if the prediction is correct or "incorrect" if the prediction is wrong.
Examples:
Question: Who authored The Taming of the Shrew (published in 2002)?
