Question Answering on MedQA (HS, Acc)
42.75HSBeavertails
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| BeavertailsModel=Qwen2.5, Mixing ratio (r)=0.12026.03 | 42.75 | 62.5 | |
| BeavertailsModel=Mistral, Mixing ratio (r)=0.12026.03 | 42.65 | 57.76 | |
| NoneModel=Mistral, Mixing ratio (r)=0.12026.03 | 41.01 | 56.93 | |
| BeavertailsModel=Llama3, Mixing ratio (r)=0.12026.03 | 31.05 | 61.56 | |
| NoneModel=Qwen2.5, Mixing ratio (r)=0.12026.03 | 15.91 | 62.27 | |
| BeavertailsModel=OLMo2, Mixing ratio (r)=0.12026.03 | 14.56 | 49.33 | |
| AegisModel=Qwen2.5, Mixing ratio (r)=0.12026.03 | 13.25 | 61.9 | |
| GR-SAPModel=Mistral, Mixing ratio (r)=0.12026.03 | 13.05 | 52.71 | |
| AegisModel=Mistral, Mixing ratio (r)=0.12026.03 | 12.51 | 55.01 | |
| GR-SAPModel=Qwen2.5, Mixing ratio (r)=0.12026.03 | 10.89 | 62.45 | |
| NoneModel=Llama3, Mixing ratio (r)=0.12026.03 | 4.78 | 61.06 | |
| AegisModel=Llama3, Mixing ratio (r)=0.12026.03 | 4.5 | 61.51 | |
| AegisModel=OLMo2, Mixing ratio (r)=0.12026.03 | 2.02 | 49.8 | |
| NoneModel=OLMo2, Mixing ratio (r)=0.12026.03 | 0.74 | 49.49 | |
| GR-SAPModel=Llama3, Mixing ratio (r)=0.12026.03 | 0.7 | 61.32 | |
| OriginalModel=OLMo2, Mixing ratio (r)=0.12026.03 | 0.54 | 50.07 | |
| GR-SAPModel=OLMo2, Mixing ratio (r)=0.12026.03 | 0.47 | 49.04 |