Generalization on LLaMA Original Languages 3.1-8B
80.37Generalization ScoreLLaMA-3.1-8B-Instruct
Evaluation Results
| Method | Links | |
|---|---|---|
| LLaMA-3.1-8B-InstructCategory=Base Model2026.05 | 80.37 | |
| DeltaMoECategory=Baseline w/ Matched Parameters2026.05 | 77.32 | |
| Dense-FT-AvgCategory=Baselines w/ Same CPT Data2026.05 | 76.54 | |
| Dense-FT-Avg-2FLOPsCategory=Baselines w/ Matched FLOPs (2x CPT Data)2026.05 | 75.78 | |
| Dense-FT-DeltaCategory=Baselines w/ Same CPT Data2026.05 | 74.37 | |
| MoLACategory=Baseline w/ Matched Parameters2026.05 | 74.21 | |
| Dense-FT-Delta-2FLOPsCategory=Baselines w/ Matched FLOPs (2x CPT Data)2026.05 | 71.62 | |
| LLaMA-ProCategory=Baseline w/ Matched Parameters2026.05 | 65.78 |