Question Answering on Math-Coding
63.46AccuracyLoRA-Mixer
Evaluation results
| Method | Links | |
|---|---|---|
| LoRA-MixerBase Model=LLaMA3-8B, Evaluator=DeepSeek-R1, trainable parameters=3.88%2025.06 | 63.46 | |
| MixLoRABase Model=LLaMA3-8B, Evaluator=DeepSeek-R1, trainable parameters=8.08%2025.06 | 63.1 | |
| MOLEBase Model=LLaMA3-8B, Evaluator=DeepSeek-R1, trainable parameters=0.04%2025.06 | 62.24 | |
| LoRAHubBase Model=LLaMA3-8B, Evaluator=DeepSeek-R12025.06 | 61.08 | |
| BaseBase Model=LLaMA3-8B, Evaluator=DeepSeek-R12025.06 | 59.37 |