VeriSci Learned Selector Joint Verifier
Experimental candidate with a learned claim-sentence selector and joint verifier.
- Selector base:
microsoft/deberta-v3-small - Verifier base:
microsoft/deberta-v3-base - SciFact validation accuracy: 0.4800
- SciFact validation macro F1: 0.3085
- Promotion gate passed:
False - Promotion gate reason:
candidate_accuracy_below_target: 0.480 < 0.910
This artifact is experimental unless the repository documentation marks it as promoted.
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support