I have a question about the Named Entity Recognition Tool.
I have be experimenting with the option to Train with New Entities.
BUT, can you turn-off or selectively turn-off the default labels?
They don't work well with domain specific terms and create noise making it difficult to focus on the initial outputs.
The "noise" is most evident when you first run a model and try to view the results to see if your custom labels are returning the desired results.