Hello, I don’t use the assistant api directly, but there is an error in the development environment configuration stage, so I leave a query.
Currently, there is a model that fine-tuned and learned about 60,000 data.
When you add a vector store to the model and you are asked to answer based on the content, a common notification is returned: “The content was not found in the document.”
This does not happen when you use a model that has been fine-tuned with other little data or a general model provided by your company.
There was no compulsory part to answer in the learning data when learning, and the environment we are currently testing is using the Assistant Tab provided by your company.
Is there a similar history or workaround for these phenomena?