Content filtering is significant in models to ensure that the generated content aligns with specific guidelines and policies. It helps prevent the generation of inappropriate, offensive, or harmful content. In Azure Open AI, content filtering capabilities are provided through Azure AI Content Safety.
To customize content filtering per your use case, follow these steps:
- Apply for modified content filters: You can submit a request via this form to apply for modified content filters.
- Once the request gets approved, customers can customize the content filters based on their use case as outlined in this documentation.
By following these steps, you can customize the content filtering behavior of Azure Open AI models to align with your application's guidelines and mitigate potential risks associated with generated content.
Please note that responsible AI practices should be considered to ensure a positive user experience while minimizing potential harms. You can refer to the provided documentation for more information on understanding and mitigating risks associated with your application.