Multiple Choice Question Answering on ConFiQA MC
91.2F1 ScoreProbeRAG
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ProbeRAGCategory=Training-Based, Backbone=Mistral-7B-v0.32025.10 | 91.2 | 89.7 | |
| ProbeRAGCategory=Training-Based, Backbone=Qwen3-8B2025.10 | 90.7 | 89.7 | |
| ProbeRAGCategory=Training-Based, Backbone=LLaMA-3.1-8B-Instruct2025.10 | 89.2 | 87.7 | |
| ProbeRAGArchitecture=Qwen2.5-7B-Instruct2025.10 | 88.8 | 86.2 | |
| CANOECategory=Training-Based, Backbone=Mistral-7B-v0.32025.10 | 87.2 | 85.7 | |
| ParamMuteCategory=Training-Based, Backbone=Qwen3-8B2025.10 | 86.3 | 80.2 | |
| ParamMuteCategory=Training-Based, Backbone=Mistral-7B-v0.32025.10 | 86.3 | 82.1 | |
| CANOECategory=Training-Based, Backbone=Qwen3-8B2025.10 | 85.2 | 81.7 | |
| CANOECategory=Training-Based, Backbone=LLaMA-3.1-8B-Instruct2025.10 | 80.9 | 74.2 | |
| ProbeRAGArchitecture=LLaMA-2-7B-Chat-HF2025.10 | 79.1 | 69.7 | |
| Context-DPOCategory=Training-Based, Backbone=LLaMA-3.1-8B-Instruct2025.10 | 76.9 | 67.7 | |
| CANOEArchitecture=LLaMA-2-7B-Chat-HF2025.10 | 73.9 | 70.2 | |
| FaithfulRAGCategory=Prompt-Based, Backbone=Qwen3-8B2025.10 | 72.3 | 60.1 | |
| Opin(Instr)Category=Prompt-Based, Backbone=Qwen3-8B2025.10 | 70.7 | 61.1 | |
| ParamMuteCategory=Training-Based, Backbone=LLaMA-3.1-8B-Instruct2025.10 | 70.6 | 60.9 | |
| FaithfulRAGCategory=Prompt-Based, Backbone=LLaMA-3.1-8B-Instruct2025.10 | 70.3 | 61.2 | |
| CANOEArchitecture=Qwen2.5-7B-Instruct2025.10 | 68.7 | 61.1 | |
| KRECategory=Prompt-Based, Backbone=LLaMA-3.1-8B-Instruct2025.10 | 68.2 | 59.8 | |
| KRECategory=Prompt-Based, Backbone=Qwen3-8B2025.10 | 67.5 | 59.1 | |
| Opin(Instr)Category=Prompt-Based, Backbone=LLaMA-3.1-8B-Instruct2025.10 | 67.4 | 57.3 | |
| COIECDCategory=Decoding-Based, Backbone=Qwen3-8B2025.10 | 66.7 | 60.8 | |
| AdaCADCategory=Decoding-Based, Backbone=Qwen3-8B2025.10 | 64.2 | 58.7 | |
| Context-DPOArchitecture=Qwen2.5-7B-Instruct2025.10 | 62.7 | 53.7 | |
| FaithfulRAGCategory=Prompt-Based, Backbone=Mistral-7B-v0.32025.10 | 60.3 | 44.3 | |
| Full-ContextCategory=Baseline, Backbone=Qwen3-8B2025.10 | 59.6 | 50.2 | |
| Context-DPOCategory=Training-Based, Backbone=Qwen3-8B2025.10 | 59.6 | 50.1 | |
| Opin(Instr)Category=Prompt-Based, Backbone=Mistral-7B-v0.32025.10 | 58.8 | 44.1 | |
| KRECategory=Prompt-Based, Backbone=Mistral-7B-v0.32025.10 | 58.7 | 45 | |
| Context-DPOArchitecture=LLaMA-2-7B-Chat-HF2025.10 | 57.9 | 32 | |
| CADCategory=Decoding-Based, Backbone=Qwen3-8B2025.10 | 57.7 | 48.3 | |
| Context-DPOCategory=Training-Based, Backbone=Mistral-7B-v0.32025.10 | 44.8 | 28.3 | |
| AdaCADCategory=Decoding-Based, Backbone=LLaMA-3.1-8B-Instruct2025.10 | 43.3 | 32.5 | |
| AdaCADCategory=Decoding-Based, Backbone=Mistral-7B-v0.32025.10 | 33.7 | 23.4 | |
| COIECDCategory=Decoding-Based, Backbone=LLaMA-3.1-8B-Instruct2025.10 | 28.5 | 24 | |
| Full-ContextCategory=Baseline, Backbone=LLaMA-3.1-8B-Instruct2025.10 | 28 | 22.5 | |
| COIECDCategory=Decoding-Based, Backbone=Mistral-7B-v0.32025.10 | 26.1 | 14.5 | |
| Full-ContextCategory=Baseline, Backbone=Mistral-7B-v0.32025.10 | 25.6 | 12.5 | |
| CADCategory=Decoding-Based, Backbone=Mistral-7B-v0.32025.10 | 16.7 | 5.9 | |
| CADCategory=Decoding-Based, Backbone=LLaMA-3.1-8B-Instruct2025.10 | 16 | 11.4 | |
| No-ContextCategory=Baseline, Backbone=Qwen3-8B2025.10 | 7.6 | 3.6 | |
| No-ContextCategory=Baseline, Backbone=LLaMA-3.1-8B-Instruct2025.10 | 5 | 2.1 | |
| No-ContextCategory=Baseline, Backbone=Mistral-7B-v0.32025.10 | 4.4 | 0.9 |