Robustness Evaluation on MMLU
88VAccDeepSeek-R1-Distill-LLaMA-8B
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| DeepSeek-R1-Distill-LLaMA-8BFramework=MASTER2025.06 | 88 | 58 | 30 | 34.09 | |
| LLaMA-3-8B-InstructFramework=MASTER2025.06 | 69 | 47 | 22 | 31.88 | |
| Gemma-2-2B-ITFramework=MASTER2025.06 | 48 | 21 | 27 | 56.25 | |
| LLaMA-3.2-1B-InstructFramework=MASTER2025.06 | 42 | 5 | 37 | 88.1 |