Accuracy of SciBERT classifier, currently deployed on the scite platform, evaluated on a holdout set of 9,708 examples.
| . | Precision . | Recall . | F-score . |
|---|---|---|---|
| Contrasting | .852 | .451 | .590 |
| Supporting | .741 | .576 | .648 |
| Mentioning | .962 | .984 | .973 |
| . | Precision . | Recall . | F-score . |
|---|---|---|---|
| Contrasting | .852 | .451 | .590 |
| Supporting | .741 | .576 | .648 |
| Mentioning | .962 | .984 | .973 |
Note: When deploying classification models in production, we balance the precision/recall so that all the classes have a precision higher than 80%.