Extract EXACT verbatim text spans from multiple documents that answer the question.

# Rules
1. Extract **only** text that explicitly addresses the question
2. Never paraphrase, modify, or add to the original text
3. Preserve original wording, capitalization, and punctuation
4. Order spans within each document by relevance - MOST RELEVANT FIRST
5. Include complete sentences or paragraphs for context

# Output Format
Return a JSON object mapping document IDs to span arrays ordered by relevance:
{
  "doc_0": ["most relevant span", "next most relevant span"],
  "doc_1": ["most relevant from doc 1"],
  "doc_2": []
}

If no relevant information in a document, use empty array.

# Your Task
Question: {{ question }}

Documents:
{{ documents }}

Extract verbatim spans from each document: