Cohere command r plus azure AI studio deploy endpoint 500's on search_queries_only request
We've recently been testing out Cohere's Command R Plus chat model via Azure AI Studio.
We're using the target endpoint from the AI Studio deployment and appending /v1/chat to the URL. It works great with a normal chat request body like so:
{
"message": "Which lego set has the greatest number of pieces?"
}
However, Cohere also offers query generation via a "search_queries_only" value in the body, but when supplied with a true value, it 500s every time with 'Internal Server Error':
{
"message": "Which lego set has the greatest number of pieces?",
"search_queries_only": true
}
To test things out further, if I supply a false value to "search_queries_only", it responds successfully just like my first json request body example.
May I get some help with this, please? This "search_queries_only" param is straight from Microsoft documention: Here