Context-faithful Question Answering on ConFiQA
13.21MRPromptCD
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| PromptCDModel=LLAMA2-7B-CHAT2026.02 | 13.21 | — | — | — | 90.42 | 13.76 | |
| PromptCDModel=MISTRALV0.3-7B-INSTRUCT2026.02 | 14.17 | — | — | — | 84.59 | 13.97 | |
| OpinModel=LLAMA2-7B-CHAT2026.02 | 14.79 | — | — | — | 88.62 | 15.38 | |
| OpinModel=MISTRALV0.3-7B-INSTRUCT2026.02 | 16.1 | — | — | — | 83.23 | 15.97 | |
| PromptCDModel=LLAMA3-8B-INSTRUCT2026.02 | 17.7 | — | — | — | 90.92 | 19.56 | |
| AttrModel=LLAMA2-7B-CHAT2026.02 | 18.61 | — | — | — | 82.03 | 18.76 | |
| AttrModel=MISTRALV0.3-7B-INSTRUCT2026.02 | 18.9 | — | — | — | 79.64 | 18.56 | |
| OpinModel=LLAMA3-8B-INSTRUCT2026.02 | 20.1 | — | — | — | 91.21 | 22.95 | |
| PromptCDModel=QWEN2.5-7B-INSTRUCT2026.02 | 24.34 | — | — | — | 91.82 | 29.54 | |
| VanillaModel=QWEN2.5-7B-INSTRUCT2026.02 | 26.05 | — | — | — | 43.78 | 15.42 | |
| AttrModel=QWEN2.5-7B-INSTRUCT2026.02 | 26.9 | — | — | — | 90.01 | 33.13 | |
| OpinModel=QWEN2.5-7B-INSTRUCT2026.02 | 27.6 | — | — | — | 89.02 | 33.93 | |
| VanillaModel=MISTRALV0.3-7B-INSTRUCT2026.02 | 28.57 | — | — | — | 64.68 | 25.87 | |
| VanillaModel=LLAMA2-7B-CHAT2026.02 | 28.78 | — | — | — | 69.66 | 28.14 | |
| VanillaModel=LLAMA3-8B-INSTRUCT2026.02 | 29.22 | — | — | — | 54.23 | 22.39 | |
| LLAMA2-7BModel Version=chat2024.12 | 29.4 | 61.5 | 25.6 | 0.5 | — | — | |
| AttrModel=LLAMA3-8B-INSTRUCT2026.02 | 29.93 | — | — | — | 87.82 | 37.52 | |
| LLAMA2-13BModel Version=chat2024.12 | 34.4 | 55.3 | 29 | 0 | — | — | |
| MISTRAL-7BModel Version=instruct-v0.22024.12 | 50.8 | 39.3 | 40.5 | 0.3 | — | — | |
| CHATGPT-42024.12 | 53.9 | 32.6 | 38.3 | 0.6 | — | — | |
| LLAMA3-8BModel Version=instruct2024.12 | 55.3 | 35.8 | 44.3 | 0 | — | — | |
| GEMINI-1.5-PRO2024.12 | 63.7 | 27.4 | 48.3 | 2.1 | — | — | |
| QWEN2-7BModel Version=instruct2024.12 | 65.4 | 24 | 43.3 | 0 | — | — | |
| CHATGPT-4o2024.12 | 82.5 | 12.1 | 56.7 | 0 | — | — |