Question Answering on SQuAD entity-level knowledge conflict (test)
9.7MRFaithfulRAG
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| FaithfulRAGBackbone LLM=mistral-7b-instruct, Category=Context-faithful Decoding2025.06 | 9.7 | 85.7 | |
| COIECDBackbone LLM=llama3.1-8b-instruct, Category=Context-faithful Decoding2025.06 | 10.7 | 71.8 | |
| FaithfulRAGBackbone LLM=llama3.1-8b-instruct, Category=Context-faithful Decoding2025.06 | 13.5 | 86.3 | |
| Opin(Instr)Backbone LLM=llama3.1-8b-instruct, Category=Context-faithful Prompting2025.06 | 13.8 | 73.4 | |
| KREBackbone LLM=llama3.1-8b-instruct, Category=Context-faithful Decoding2025.06 | 14.5 | 66.1 | |
| CADBackbone LLM=llama3.1-8b-instruct, Category=Context-faithful Decoding2025.06 | 14.6 | 71.2 | |
| COIECDBackbone LLM=qwen2.5-7b-instruct, Category=Context-faithful Decoding2025.06 | 15.5 | 70.8 | |
| FaithfulRAGBackbone LLM=qwen2.5-7b-instruct, Category=Context-faithful Decoding2025.06 | 15.7 | 78.3 | |
| COIECDBackbone LLM=mistral-7b-instruct, Category=Context-faithful Decoding2025.06 | 16.1 | 65.4 | |
| ATTRBackbone LLM=llama3.1-8b-instruct, Category=Context-faithful Prompting2025.06 | 16.8 | 69.5 | |
| KREBackbone LLM=mistral-7b-instruct, Category=Context-faithful Decoding2025.06 | 16.8 | 74.6 | |
| CADBackbone LLM=mistral-7b-instruct, Category=Context-faithful Decoding2025.06 | 18 | 66.9 | |
| ChatQA-1.5Backbone LLM=Llama3.1-8B, Category=Specific RAG Models2025.06 | 19.1 | 77 | |
| ATTRBackbone LLM=mistral-7b-instruct, Category=Context-faithful Prompting2025.06 | 20.6 | 70.2 | |
| OriginBackbone LLM=llama3.1-8b-instruct, Category=With Full Context2025.06 | 20.7 | 69.5 | |
| ChatQA-2.0Backbone LLM=Llama3.1-8B, Category=Specific RAG Models2025.06 | 20.8 | 75.4 | |
| CADBackbone LLM=qwen2.5-7b-instruct, Category=Context-faithful Decoding2025.06 | 21.7 | 73.4 | |
| ATTRBackbone LLM=qwen2.5-7b-instruct, Category=Context-faithful Prompting2025.06 | 22.2 | 72.9 | |
| KREBackbone LLM=qwen2.5-7b-instruct, Category=Context-faithful Decoding2025.06 | 22.7 | 73.7 | |
| Opin(Instr)Backbone LLM=mistral-7b-instruct, Category=Context-faithful Prompting2025.06 | 22.9 | 69.3 | |
| OriginBackbone LLM=mistral-7b-instruct, Category=With Full Context2025.06 | 25 | 67.2 | |
| Opin(Instr)Backbone LLM=qwen2.5-7b-instruct, Category=Context-faithful Prompting2025.06 | 25.5 | 70.5 | |
| OriginBackbone LLM=qwen2.5-7b-instruct, Category=With Full Context2025.06 | 27.6 | 68.3 | |
| Self-RAGBackbone LLM=Llama2-7B, Category=Specific RAG Models2025.06 | 30.1 | 62 |