General Performance on MMLU, HellaSwag, TruthfulQA, GSM8K, MATH, MBPP, HumanEval
40.35Average ScoreSens-Merging (DARE)
Evaluation Results
| Method | Links | |
|---|---|---|
| Sens-Merging (DARE)Backbone=LLaMA2-13B, Use Sens=true2025.02 | 40.35 | |
| Sens-Merging (Ties-Merging)Backbone=LLaMA2-13B, Use Sens=true2025.02 | 40.22 | |
| Sens-Merging (Task Arithmetic)Backbone=LLaMA2-13B, Use Sens=true2025.02 | 40.2 | |
| DAREBackbone=LLaMA2-13B, Use Sens=false2025.02 | 40.13 | |
| Ties-MergingBackbone=LLaMA2-13B, Use Sens=false2025.02 | 39.89 | |
| MathBackbone=LLaMA2-13B2025.02 | 37.42 | |
| Task ArithmeticBackbone=LLaMA2-13B, Use Sens=false2025.02 | 34.52 | |
| CodeBackbone=LLaMA2-13B2025.02 | 31.21 | |
| ChatBackbone=LLaMA2-13B2025.02 | 31.18 |