Policy-referencing outputs

GPT
Template Prompt
Intermediate

Prompt

If the content violates this policy, return:
{"violation": 1, "policy_category": "<category_or_rule_id>"}

If the content does NOT violate this policy, return:
{"violation": 0, "policy_category": null}

Example:
{"violation": 1, "policy_category": "H2.f"}