Generate Ground Truth Responses For Questions Based On Contexts Found In The Document

You have generated a list of questions and their contexts, based on a document that was provided to you. Each question is associated with relevant context from the document.

Instructions:

1. Analyze: Carefully read and understand the questions thoroughly. Examine each provided context carefully, identifying key information and details that pertain to the question.

2. Analyze: Synthesize the identified information from all questions and contexts to generate well-supported ground truth answers in {language}.

3. Generate answers: Provide the answer to the question in {language} using information exclusively from the provided context. The answer should be precise and should directly address the question based on its context.

4. Output Format: Please output the questions and their corresponding ground truth answers in {language} formatted as a JSON dictionary. This dictionary should consist of an array of objects under the key "questions", where each object contains two properties: "question" and "ground_truth".

Example Format: "questions": "question": "str", "ground_truth": "str"

Example of Existing Output: "questions": "question": "How often does prosperity adjustment take place?", "ground_truth": "Every year the social partners and the government distribute a welfare envelope."

Omit any supplementary information or redundant text. Do not answer with OUTPUT

{questions}