Jev in production › Evaluation and testing
Benchmarks Jev and rival models on 7,500 questions scored against human annotator disagreement from NVIDIA HelpSteer2, rather than one correct answer (trust123500).
Open on github.com ↗Copy link
For the project's own README, linking back here: