| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| MMMLU Danish (test) | TV + KMM | MMLU Score46.24 | 25 | 15d ago | |
| Heterogeneous SFT Benchmarks (Evaluation) | EPI | GSM8K Score65 | 20 | 1mo ago | |
| Marathi Multilingual SFT (test) | TV+KMM | Best-k Accuracy34.76 | 9 | 15d ago | |
| Danish Multilingual SFT (test) | TV+KMM | Best-k Accuracy46.38 | 9 | 15d ago | |
| MMMLU Marathi (test) | TV + KMM | MMMLU Accuracy34.6 | 9 | 15d ago | |
| SFT (evaluation) | BHyT | SFT Evaluation Loss3.13 | 5 | 3mo ago | |
| SFT (train) | Peri-LN | SFT Train Loss2.614 | 5 | 3mo ago | |
| Supervised Fine-Tuning (SFT) | BHyT | SFT Training Loss2.468 | 2 | 3mo ago |
