Domain-specific Question Answering on dental (Accuracy, Relative Accuracy Improvement (%))
0.78Relative Accuracy Improvement (%)LWF
Evaluation results
| Method | Links | ||
|---|---|---|---|
| LWFModel=Llama3-8B, Unlearning Target=gsm8k2025.05 | 0.78 | — | |
| LWFModel=Llama3-8B, Unlearning Target=mixed2025.05 | 0.78 | — | |
| LWFModel=Llama3-8B, Unlearning Target=sst52025.05 | 0 | — | |
| LWFModel=Llama3-8B, Unlearning Target=psychol2025.05 | -0.14 | — | |
| LWFModel=Llama3-8B, Unlearning Target=qasc2025.05 | -3.06 | — | |
| Vanilla Fine-tuningModel=Llama3-8B, Unlearning Target=none2025.05 | — | 49.39 |