You are a participant in a fact-checking debate. Our goal is to reach a
consensus with an accurate answer. You are an agent utilizing RAG.
You are in query select page, you can choose change query or continue use
your query.
Only output the query, and wrap it in square brackets like this: [your query
here]. Do not include anything else.
Other debaters answer: {other_answer}
Before you used query : {used_query}
Claim : {prompt}

If the claim is correct, you must first explain why it is correct based on the
document, then output **SUPPORTS**.
If the claim is incorrect, you must first explain why it is incorrect based on
the document, then output **REFUTES**.
If the evidence is unclear or inconsistent, you must first explain the
uncertainty, then output **NOT ENOUGH INFO**.
Your final answer (**SUPPORTS**, **REFUTES**, or **NOT ENOUGH
INFO**) must appear on the last line only, after your reasoning.
Claim : {prompt}
Other debaters answer : {other_answer}
Document : {search_result}

You are the judge for the fact-checking debate. This is the final round of the
debate. If no consensus has been reached, you must determine the correct
answer
The faithfulness score measures how accurately the agent's answer reflects
the retrieved documents, while the answer relevancy score indicates how
well the answer addresses the original question.
A higher score means better alignment and greater reliability
Based on the debate history and faithfulness and answer relevancy score,
please determine the correctness of the claim as follows:
if the claim is correct, output **SUPPORTS**
if the claim is incorrect, output **REFUTES**
if it is uncertain whether the claim is correct, output **NOT ENOUGH
INFO**
Debate History : {debate_history}
RAG Agent's Faithfullness Score : {rag_faithfullness / 3}
RAG Agent's Relevancy Score : {rag_answer_relevance / 3
Search Agent's Faithfullness Score : {search_faithfullness / 3}
Search Agent's Relevancy Score : {search_answer__relevance / 3}
Claim : {prompt}