Question Answering on SQuAD (Accuracy (ACC))
86.3Accuracy (ACC)FaithfulRAG
Evaluation Results
| Method | Links | |
|---|---|---|
| FaithfulRAGBackbone LLM=llama3.1-8b-instruct2025.06 | 86.3 | |
| FaithfulRAGBackbone LLM=mistral-7b-instruct2025.06 | 85.7 | |
| FaithfulRAGBackbone LLM=qwen2.5-7b-instruct2025.06 | 78.3 | |
| ChatQA-1.5Backbone LLM=Llama3.1-8B2025.06 | 77 | |
| ChatQA-2.0Backbone LLM=Llama3.1-8B2025.06 | 75.4 | |
| KREBackbone LLM=mistral-7b-instruct2025.06 | 74.6 | |
| KREBackbone LLM=qwen2.5-7b-instruct2025.06 | 73.7 | |
| Opin(Instr)Backbone LLM=llama3.1-8b-instruct2025.06 | 73.4 | |
| CADBackbone LLM=qwen2.5-7b-instruct2025.06 | 73.4 | |
| ATTRBackbone LLM=qwen2.5-7b-instruct2025.06 | 72.9 | |
| COIECDBackbone LLM=llama3.1-8b-instruct2025.06 | 71.8 | |
| CADBackbone LLM=llama3.1-8b-instruct2025.06 | 71.2 | |
| COIECDBackbone LLM=qwen2.5-7b-instruct2025.06 | 70.8 | |
| Opin(Instr)Backbone LLM=qwen2.5-7b-instruct2025.06 | 70.5 | |
| ATTRBackbone LLM=mistral-7b-instruct2025.06 | 70.2 | |
| Origin model with full contextBackbone LLM=llama3.1-8b-instruct2025.06 | 69.5 | |
| ATTRBackbone LLM=llama3.1-8b-instruct2025.06 | 69.5 | |
| Opin(Instr)Backbone LLM=mistral-7b-instruct2025.06 | 69.3 | |
| Origin model with full contextBackbone LLM=qwen2.5-7b-instruct2025.06 | 68.3 | |
| Origin model with full contextBackbone LLM=mistral-7b-instruct2025.06 | 67.2 | |
| CADBackbone LLM=mistral-7b-instruct2025.06 | 66.9 | |
| KREBackbone LLM=llama3.1-8b-instruct2025.06 | 66.1 | |
| COIECDBackbone LLM=mistral-7b-instruct2025.06 | 65.4 | |
| Self-RAGBackbone LLM=Llama2-7B2025.06 | 62 | |
| Origin model without contextBackbone LLM=mistral-7b-instruct2025.06 | 11.5 | |
| Origin model without contextBackbone LLM=llama3.1-8b-instruct2025.06 | 11.2 | |
| Origin model without contextBackbone LLM=qwen2.5-7b-instruct2025.06 | 11.1 |