Science Completion on HellaSwag
95.2AccuracyMixDoRA
Evaluation results
| Method | Links | |
|---|---|---|
| MixDoRABase Model=LLaMA-2 13B, Trainable Parameters=2.5%2024.04 | 95.2 | |
| DoRABase Model=LLaMA-2 13B, Trainable Parameters=2.4%2024.04 | 94.8 | |
| MixLoRABase Model=LLaMA-2 13B, Trainable Parameters=2.5%2024.04 | 94.7 | |
| MixDoRABase Model=LLaMA-3 8B, Trainable Parameters=3.0%2024.04 | 94.6 | |
| LoRABase Model=LLaMA-2 13B, Trainable Parameters=2.4%2024.04 | 94.3 | |
| MixLoRABase Model=LLaMA-3 8B, Trainable Parameters=3.0%2024.04 | 93.3 | |
| MixLoRABase Model=LLaMA-2 7B, Trainable Parameters=2.9%2024.04 | 93.1 | |
| MixDoRABase Model=LLaMA-2 7B, Trainable Parameters=2.9%2024.04 | 90.6 | |
| DoRABase Model=LLaMA-2 7B, Trainable Parameters=2.9%2024.04 | 89.6 | |
| MixDoRABase Model=Gemma 2B, Trainable Parameters=4.3%2024.04 | 89.3 | |
| MixLoRABase Model=Gemma 2B, Trainable Parameters=4.3%2024.04 | 89 | |
| LoRABase Model=LLaMA-2 7B, Trainable Parameters=2.9%2024.04 | 88.4 | |
| DoRABase Model=Gemma 2B, Trainable Parameters=3.2%2024.04 | 85.4 | |
| LoRABase Model=Gemma 2B, Trainable Parameters=3.2%2024.04 | 84.4 | |
| DoRABase Model=LLaMA-3 8B, Trainable Parameters=2.6%2024.04 | 78.8 | |
| LoRABase Model=LLaMA-3 8B, Trainable Parameters=2.6%2024.04 | 74.2 |