Personal Question Answering on PERQA (test)
38.6Hit@1REQAP
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| REQAPModel=GPT4o, Model Scale=>100B2025.05 | 38.6 | 52 | |
| REQAPModel=SFT, Model Scale=1B2025.05 | 38 | 53 | |
| REQAPModel=LLaMA3.3, Model Scale=70B2025.05 | 32.2 | 46 | |
| CODEGENModel=GPT4o, Model Scale=>100B2025.05 | 31.9 | 44 | |
| CODEGENModel=SFT, Model Scale=1B2025.05 | 31.5 | 47 | |
| CODEGENModel=LLaMA3.3, Model Scale=70B2025.05 | 23.9 | 33 | |
| RAGModel=GPT4o, Model Scale=>100B2025.05 | 14.9 | 20 | |
| RAGModel=LLaMA3.3, Model Scale=70B2025.05 | 12.3 | 18 | |
| RAGModel=SFT, Model Scale=1B2025.05 | 2.9 | 6 |