Question Answering on UNQOVER
99.75AccuracyCAL
Evaluation Results
| Method | Links | |
|---|---|---|
| CALBackbone=Llama3-8B, Setting=Zero-shot2025.04 | 99.75 | |
| CALBackbone=Llama3-8B, Setting=Few-shot2025.04 | 95.67 | |
| FairSteerBackbone=Llama3-8B, Setting=Few-shot2025.04 | 91.94 | |
| CALBackbone=Mistral-7B, Setting=Few-shot2025.04 | 87.02 | |
| Llama3-8BSetting=Few-shot2025.04 | 76.22 | |
| FairSteerBackbone=Vicuna-13B, Setting=Few-shot2025.04 | 73.19 | |
| FairSteerBackbone=Mistral-7B, Setting=Few-shot2025.04 | 69.51 | |
| CALBackbone=Mistral-7B, Setting=Zero-shot2025.04 | 67.54 | |
| CALBackbone=Llama2-7B, Setting=Zero-shot2025.04 | 62.52 | |
| CALBackbone=Llama2-13B, Setting=Zero-shot2025.04 | 60.32 | |
| CALBackbone=Vicuna-13B, Setting=Few-shot2025.04 | 58.76 | |
| FairSteerBackbone=Llama3-8B, Setting=Zero-shot2025.04 | 58.01 | |
| Mistral-7BSetting=Few-shot2025.04 | 57.83 | |
| FairSteerBackbone=Vicuna-7B, Setting=Few-shot2025.04 | 57.21 | |
| FairSteerBackbone=Mistral-7B, Setting=Zero-shot2025.04 | 54.1 | |
| FairSteerBackbone=Llama2-13B, Setting=Zero-shot2025.04 | 53 | |
| Vicuna-13BSetting=Few-shot2025.04 | 52.9 | |
| FairSteerBackbone=Llama2-13B, Setting=Few-shot2025.04 | 49.23 | |
| FairSteerBackbone=Vicuna-13B, Setting=Zero-shot2025.04 | 49.06 | |
| FairSteerBackbone=Vicuna-7B, Setting=Zero-shot2025.04 | 43.11 | |
| Vicuna-13BSetting=Zero-shot2025.04 | 41.44 | |
| CALBackbone=Vicuna-13B, Setting=Zero-shot2025.04 | 35.34 | |
| Llama2-13BSetting=Zero-shot2025.04 | 33.96 | |
| CALBackbone=Vicuna-7B, Setting=Zero-shot2025.04 | 33.66 | |
| Mistral-7BSetting=Zero-shot2025.04 | 33.06 | |
| CALBackbone=Llama2-13B, Setting=Few-shot2025.04 | 32.27 | |
| FairSteerBackbone=Llama2-7B, Setting=Few-shot2025.04 | 28.52 | |
| Llama3-8BSetting=Zero-shot2025.04 | 20.84 | |
| Llama2-7BSetting=Few-shot2025.04 | 20.6 | |
| FairSteerBackbone=Llama2-7B, Setting=Zero-shot2025.04 | 19.67 | |
| Llama2-13BSetting=Few-shot2025.04 | 19.17 | |
| Vicuna-7BSetting=Few-shot2025.04 | 18.34 | |
| Vicuna-7BSetting=Zero-shot2025.04 | 16.19 | |
| CALBackbone=Llama2-7B, Setting=Few-shot2025.04 | 12.3 | |
| CALBackbone=Vicuna-7B, Setting=Few-shot2025.04 | 10.88 | |
| Llama2-7BSetting=Zero-shot2025.04 | 7.03 |