CaseHOLD
CaseHOLD: Measures legal reasoning, contract review, statute interpretation, or legal-domain QA.
8rows
micro_macro_f1primary metric
2026-05-27sampled
Metadata
Metrics
Micro/Macro F1
| Rank | Subject | Micro/Macro F1 | Model Match | Provenance | Sampled |
|---|---|---|---|---|---|
| 1 | CaseLaw-BERT | 75.4 | — | Imported | 2026-05-27 |
| 2 | Legal-BERT | 75.3 | — | Imported | 2026-05-27 |
| 3 | DeBERTa | 72.6 | — | Imported | 2026-05-27 |
| 4 | Longformer | 71.9 | — | Imported | 2026-05-27 |
| 5 | RoBERTa | 71.4 | — | Imported | 2026-05-27 |
| 6 | BERT | 70.8 | — | Imported | 2026-05-27 |
| 7 | BigBird | 70.8 | — | Imported | 2026-05-27 |
| 8 | TFIDF-SVM | 22.4 | — | Imported | 2026-05-27 |
No matching rows.