Prompt example extracted from OpenAI Cookbook.
EXAMPLES Provide 4–6 short examples labeled 0 or 1. Content: [INPUT] Answer (0 or 1): ``` To reduce the likelihood of false positives or confusion, avoid using words like “generally” or “usually”. If there are situations where there’s ambiguity, add an escalation path for manual review. This is also especially helpful for regional or language differences. Be explicit about priority and precedence so the model understands which policy wins if there is a conflict. If there are multiple policy violations, define which one is dominant.
