General Knowledge on MMLU, HellaSwag, TruthfulQA
55.88MMLUSens-Merging (Task Arithmetic)
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Sens-Merging (Task Arithmetic)Backbone=LLaMA2-13B, Use Sens=true2025.02 | 55.88 | 61.84 | 39.05 | |
| Sens-Merging (DARE)Backbone=LLaMA2-13B, Use Sens=true2025.02 | 55.65 | 61.66 | 40.64 | |
| Ties-MergingBackbone=LLaMA2-13B, Use Sens=false2025.02 | 55.48 | 60.65 | 39.05 | |
| DAREBackbone=LLaMA2-13B, Use Sens=false2025.02 | 55.43 | 61.51 | 40.51 | |
| Sens-Merging (Ties-Merging)Backbone=LLaMA2-13B, Use Sens=true2025.02 | 55.2 | 60.64 | 39.17 | |
| ChatBackbone=LLaMA2-13B2025.02 | 53.17 | 60.73 | 40.88 | |
| MathBackbone=LLaMA2-13B2025.02 | 52.73 | 61.1 | 37.09 | |
| CodeBackbone=LLaMA2-13B2025.02 | 52.65 | 60.42 | 40.64 | |
| Task ArithmeticBackbone=LLaMA2-13B, Use Sens=false2025.02 | 52.22 | 57.52 | 41.49 |