Claim Tagging on SubstanReview 110 reviews (evaluation)
61.36PrecisionModernBERT (Our)
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| ModernBERT (Our)Backbone=ModernBERT-large, Maximum input length=8096, Epochs=20, Batch size=8, Learning rate=4e-52026.01 | 61.36 | 53.13 | 56.91 | |
| Guo et al.2026.01 | 52 | 59.77 | 55.61 |