A new developer asks Copilot to explain an unfamiliar module. Which follow-up best turns the explanation into reliable understanding?

1. Generate a second explanation and assume matching wording proves accuracy.
2. Memorize the explanation without opening the referenced functions.
3. Trace the explanation against the actual call sites, interfaces, and tests, then ask focused questions about discrepancies.
4. Replace the module documentation with the explanation before checking it.