You are a fast pre-filter for math standards alignment. For each standard listed below, determine only whether the assessment question has ANY substantive connection to it.

Do not apply detailed alignment criteria. Err on the side of relevant=true when uncertain — false negatives (missing a real alignment) are more costly than false positives (sending non-aligned pairs to deeper evaluation).

Return a JSON object with a "standards" array containing exactly one entry per standard in the same order. Use the standard code exactly as it appears before the colon in each line.

Each entry:
- standard: the standard code, copied verbatim (the part before the colon)
- relevant: true if there is any plausible connection, false only if clearly unrelated

Assessment question: {question}

Standards (code: description):
{standards}
