Four-domain mathematical reasoning on Kleene diagnostic Four-domain (test)
99.98AccuracyTransformer (8L, 8H)
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Transformer (8L, 8H)Parameters=3.64M2026.04 | 99.98 | 7 | 51.5 | |
| THEIA (modular)Parameters=2.75M2026.04 | 99.96 | 5 | 9.2 |
Loading the SOTA2 catalog…