Added native JSON Structured formatting to the generated response
Added a boolean flag to control the sampling of the full or partial dataset. For partial sampling, added a if-else to add the eval samples in the vectorstore to ensure parity between ground-truth and eval datasets.
Fixed issues with eval data persistence
Fixed histogram plot at the end
End-to-end notebook run successfully complete before final push.
This PR covers the following changes:
End-to-end notebook run successfully complete before final push.