Multitask Language Understanding on MMMLU (target) reporting RPR and WPR
57.86RPRORPO
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ORPOMode=Fine-tuned, English occurrence=language confusion2026.04 | 57.86 | 71.43 | |
| DPOMode=Fine-tuned, English occurrence=language confusion2026.04 | 55.64 | 65.44 | |
| TLPON=16, Mode=Fine-tuned, English occurrence=language confusion2026.04 | 52.25 | 60.33 | |
| SFTMode=Fine-tuned, English occurrence=language confusion2026.04 | 44.61 | 55.2 | |
| BaselineDescription=Original model prior to fine-tuning2026.04 | 43.66 | 58.26 |