Question Answering on MuSiQue golden (Accuracy)
85.3AccuracyFaithfulRAG
Evaluation Results
| Method | Links | |
|---|---|---|
| FaithfulRAGBackbone=LLaMA 3.1-8B-Instruct2025.06 | 85.3 | |
| OriginBackbone=LLaMA 3.1-8B-Instruct2025.06 | 84 | |
| COIECDBackbone=LLaMA 3.1-8B-Instruct, Method Category=Context-faithful Decoding2025.06 | 83.3 | |
| Opin (Instr)Backbone=LLaMA 3.1-8B-Instruct, Method Category=Context-faithful Prompting2025.06 | 83 | |
| CADBackbone=LLaMA 3.1-8B-Instruct, Method Category=Context-faithful Decoding2025.06 | 81.9 | |
| ATTRBackbone=LLaMA 3.1-8B-Instruct, Method Category=Context-faithful Prompting2025.06 | 80.4 | |
| KREBackbone=LLaMA 3.1-8B-Instruct, Method Category=Context-faithful Prompting, note=Sharp decline in performance as model refuses to generate responses2025.06 | 38.3 |