Instruction Following on MIF (target)
50.47AccuracyBaseline
Evaluation results
| Method | Links | |
|---|---|---|
| BaselineFine-tuning=Prior to fine-tuning2026.04 | 50.47 | |
| BaselineN=16, setting=English occurrence as language confusion, aggregation=Averaged across four models and four target languages2026.04 | 50.37 | |
| TLPOFine-tuning=After fine-tuning2026.04 | 49.71 | |
| DPOFine-tuning=After fine-tuning2026.04 | 46.68 | |
| TLPON=16, setting=English occurrence as language confusion, aggregation=Averaged across four models and four target languages2026.04 | 46.21 | |
| ORPOFine-tuning=After fine-tuning2026.04 | 43.82 | |
| DPON=16, setting=English occurrence as language confusion, aggregation=Averaged across four models and four target languages2026.04 | 43.77 | |
| ORPON=16, setting=English occurrence as language confusion, aggregation=Averaged across four models and four target languages2026.04 | 42.62 | |
| SFTFine-tuning=After fine-tuning2026.04 | 39.91 | |
| SFTN=16, setting=English occurrence as language confusion, aggregation=Averaged across four models and four target languages2026.04 | 39.91 |