Why does the prompt "give me top 3 services" for 'gpt-35-turbo-16k' trigger Azure OpenAI's content management policy 'Sexual (medium)'

Question

0

We are working on a NLP-to-SQL POC. While submitting a prompt to Azure Open AI we are facing a strange error. We have narrowed it down to the prompt 'give me top 3 services'. Please see the screenshot below:

When I submit the prompt 'give me top 3 services' to the standard deployment of 'gpt-35-turbo-16k' in Azure, it gives me an error saying:

"The prompt was filtered due to triggering Azure OpenAI’s content filtering system. Reason: This prompt contains content flagged as Sexual (medium)"

Don't understand why it's giving me this error and how to solve it?

When I ask "is there anything offensive or sexual in 'give me top 3 services'" all I get back is:

"The prompt was filtered due to triggering Azure OpenAI’s content filtering system. Reason: This prompt contains content flagged as Sexual (low)"

Answer 1

Jignesh Vaducha Greetings & Welcome to Microsoft Q&A forum!

"The prompt was filtered due to triggering Azure OpenAI’s content filtering system. Reason: This prompt contains content flagged as Sexual (low)"

You can try rephrasing the prompt to remove any potentially offensive or inappropriate content.

Try to provide more details like what specific services you would like to retrieve?

For example, you could try "What are the top 3 services in Azure?" or "Which services are the most popular in Azure?"

Please see Content Filter Configurability to more about Content Filtering.

Do let me know if that helps or have any further queries.