Open-domain Question Answering on NQ-Open
29AccuracyBaseline
Evaluation Results
| Method | Links | |
|---|---|---|
| BaselineParam Ratio=1.00, Base Model=LLaMA-3-8b, Fine-tuning Protocol=None2025.12 | 29 | |
| BaselineParam Ratio=1.00, Base Model=LLaMA-2-7b, Fine-tuning Protocol=None2025.12 | 26 | |
| LLRCParam Ratio=0.90, Base Model=LLaMA-2-7b, Fine-tuning Protocol=None2025.12 | 22.1 | |
| LLRCParam Ratio=0.90, Base Model=LLaMA-3-8b, Fine-tuning Protocol=None2025.12 | 19.6 | |
| LLM-Pruner+FinetuneParam Ratio=0.90, Base Model=LLaMA-3-8b, Fine-tuning Protocol=Alpaca dataset2025.12 | 19 | |
| LLRCParam Ratio=0.85, Base Model=LLaMA-2-7b, Fine-tuning Protocol=None2025.12 | 18.5 | |
| LLM-Pruner+FinetuneParam Ratio=0.90, Base Model=LLaMA-2-7b, Fine-tuning Protocol=Alpaca dataset2025.12 | 17.9 | |
| LLM-PrunerParam Ratio=0.90, Base Model=LLaMA-3-8b, Fine-tuning Protocol=None2025.12 | 17 | |
| LLM-PrunerParam Ratio=0.90, Base Model=LLaMA-2-7b, Fine-tuning Protocol=None2025.12 | 15.5 | |
| LLM-Pruner+FinetuneParam Ratio=0.85, Base Model=LLaMA-2-7b, Fine-tuning Protocol=Alpaca dataset2025.12 | 14.7 | |
| LLM-Pruner+FinetuneParam Ratio=0.85, Base Model=LLaMA-3-8b, Fine-tuning Protocol=Alpaca dataset2025.12 | 14.7 | |
| LLRCParam Ratio=0.85, Base Model=LLaMA-3-8b, Fine-tuning Protocol=None2025.12 | 14 | |
| LLRCParam Ratio=0.80, Base Model=LLaMA-2-7b, Fine-tuning Protocol=None2025.12 | 13.5 | |
| LLM-PrunerParam Ratio=0.85, Base Model=LLaMA-2-7b, Fine-tuning Protocol=None2025.12 | 13.1 | |
| LLM-PrunerParam Ratio=0.85, Base Model=LLaMA-3-8b, Fine-tuning Protocol=None2025.12 | 11.8 | |
| LLM-Pruner+FinetuneParam Ratio=0.80, Base Model=LLaMA-2-7b, Fine-tuning Protocol=Alpaca dataset2025.12 | 11.1 | |
| LLM-Pruner+FinetuneParam Ratio=0.80, Base Model=LLaMA-3-8b, Fine-tuning Protocol=Alpaca dataset2025.12 | 10.6 | |
| LLRCParam Ratio=0.80, Base Model=LLaMA-3-8b, Fine-tuning Protocol=None2025.12 | 8.6 | |
| LLM-PrunerParam Ratio=0.80, Base Model=LLaMA-2-7b, Fine-tuning Protocol=None2025.12 | 7.8 | |
| LLM-PrunerParam Ratio=0.80, Base Model=LLaMA-3-8b, Fine-tuning Protocol=None2025.12 | 5.5 |