<s>[INST]
Your role is to evaluate the correctness of the candidate model output provided in the candidate model output section
based on whether it aligns with the ground truth answer provided in the ground truth section in answering the 
question in the question section.

Refer to the question that you have to use while evaluating the correctness of the candidate model response in alignment to the ground truth:
Question: {question}


Refer to the candidate model response to be evaluated below:
candidate model response: {answer}


Refer to the ground truth to the question in the context below in the ground_truth section:
ground_truth: {ground_truth}

Follow the instructions below while giving your evaluation in the evaluation_instructions section below:

evaluation_instructions:
{rules}

Only give your response in a JSON format. Do not prefix your response with ``` json .....  ``` and so on. 

Do not add anything else in your response. Give your response only in a correct formatted json starting with an 
opening and ending with a bracket. there should be no `````` in your response. your response
should only and only contain the json without any additional characters. Your response
should only be the JSON because it will be parsed later for downstream tasks. If you
provide anything other than or in addition to the JSON response, it will throw errors 
while performing downstream tasks.

Do not add any pre filler words. Your response should contain the JSON structured output only.

[/INST]
Answer: