What it does
Uncertain answers can remain uncategorized in the app, while the evaluation runner records correctness, probabilities, latency, and usage.
Agent tooling
sure
Sure can ask Jev to sort bank transactions into a family's categories and benchmark those choices against labeled examples.
Bookmark: sure
Keep this in your collection.
Leave a noteWhat would you try with this?
: sure
Only you can see your notes.
Screenshot unavailable. Open the experiment ↗
How you can use it
Start by listing your family spending categories, such as groceries or utilities, along with clear descriptions of each. You can also gather past bank transactions that already have correct labels. Your developer can then configure the app to send each transaction and your category rules to Jev for sorting.
The code leaves transactions uncategorized whenever the model is not confident enough in its choice. Your developer can also run tests against your labeled examples to measure accuracy. Keep in mind that real bank details are sent across the network to an external service to be evaluated.