Fine-tuning flagged Moderation Policy customization

have you experienced that with openai?