We call verbatims that a user has applied labels or entities to: ‘reviewed’. In the platform, we show which verbatims have been reviewed by styling them as seen in the examples below.
For verbatims with assigned labels, the labels will become a darker shade, two ticks will appear in front of the labels and a grey box will surround them.
For verbatims with assigned entities, a bold coloured box will surround the assigned entity and the paragraph the entity is in will be shaded in grey.
Example verbatim with assigned labels
Example verbatim with assigned entities
Example verbatim with assigned labels and entities
Unreviewed verbatims are all of the verbatims that have not been manually reviewed and labelled by a user. They will likely have predicted labels (and entities if enabled) associated with them, with various confidence levels, but these predictions have not been confirmed by a human.
Example unreviewed verbatim from a dataset with sentiment analysis disabled
Example unreviewed verbatim from a dataset with sentiment analysis enabled