Your goal is to check if one of the Predicted Insights extracted from a report is a Golden Insight. You must be STRICT and pay attention to every small detail.

Instructions:
* Evaluate if the Predicted Insights contain sufficient information to derive a Golden Insight.
* Select the insight that most closely matches the Golden Insight. Select one and only one insight.
* Answer of yes or no where:
    - yes: Selected insight contains comprehensive information to fully derive the expected insight
    - no: Selected insight lacks the necessary information, misses key details, or has significant gaps 
* Be STRICT - do not answer yes for partial matches, vague similarities, or general information. However, no exact wording is required and paraphrasing is acceptable. 
* IMPORTANT: Only consider details given in the Golden Insight when answering yes or no. Don't expect anything more than what is given in the Golden Insight.
* Focus on factual accuracy, completeness, and specificity.

Predicted Insights:
{claims_text}

Golden Insight:
{gold_insight}

Return a valid json dictionary with the following structure:
{{
    "answer": "the answer yes or no as described above",
    "confidence": "how confident you are that your answer is correct", 
    "justification": "a detailed justification explaining exactly what information given in the golden insight is present or missing in the selected insight",
    "selected_insight": "the claim text of the selected insight that most closely matches the golden insight, but not the number of the selected insight",
    "golden_insight": {gold_insight}
}}

Ensure only a json dictionary is returned, and return nothing else.