Open-ended Question Answering on ConFiQA (test)
95.7F1 ScoreProbeRAG
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ProbeRAGCategory=Training-Based, Backbone=Qwen3-8B2025.10 | 95.7 | 94.3 | |
| ProbeRAGCategory=Training-Based, Backbone=Mistral-7B-v0.32025.10 | 95.1 | 93.7 | |
| ProbeRAGCategory=Training-Based, Backbone=LLaMA-3.1-8B-Instruct2025.10 | 93.1 | 91.7 | |
| ParamMuteCategory=Training-Based, Backbone=Mistral-7B-v0.32025.10 | 93.1 | 91.2 | |
| CANOECategory=Training-Based, Backbone=Mistral-7B-v0.32025.10 | 92.5 | 90.7 | |
| CANOECategory=Training-Based, Backbone=Qwen3-8B2025.10 | 92.2 | 86.5 | |
| ParamMuteCategory=Training-Based, Backbone=Qwen3-8B2025.10 | 90.5 | 83.7 | |
| KRECategory=Prompt-Based, Backbone=Mistral-7B-v0.32025.10 | 84.5 | 72.8 | |
| KRECategory=Prompt-Based, Backbone=LLaMA-3.1-8B-Instruct2025.10 | 84 | 74.7 | |
| Context-DPOCategory=Training-Based, Backbone=LLaMA-3.1-8B-Instruct2025.10 | 83.7 | 76.7 | |
| CANOECategory=Training-Based, Backbone=LLaMA-3.1-8B-Instruct2025.10 | 82.3 | 77.7 | |
| KRECategory=Prompt-Based, Backbone=Qwen3-8B2025.10 | 80.4 | 67.3 | |
| Opin(Instr)Category=Prompt-Based, Backbone=Qwen3-8B2025.10 | 78.8 | 69.2 | |
| COIECDCategory=Decoding-Based, Backbone=Qwen3-8B2025.10 | 78.5 | 73.6 | |
| ParamMuteCategory=Training-Based, Backbone=LLaMA-3.1-8B-Instruct2025.10 | 78.3 | 73.2 | |
| Opin(Instr)Category=Prompt-Based, Backbone=LLaMA-3.1-8B-Instruct2025.10 | 76.9 | 67.4 | |
| AdaCADCategory=Decoding-Based, Backbone=Qwen3-8B2025.10 | 76.9 | 68.3 | |
| Opin(Instr)Category=Prompt-Based, Backbone=Mistral-7B-v0.32025.10 | 76.4 | 65.5 | |
| FaithfulRAGCategory=Prompt-Based, Backbone=Qwen3-8B2025.10 | 75.3 | 65.3 | |
| FaithfulRAGCategory=Prompt-Based, Backbone=Mistral-7B-v0.32025.10 | 74.2 | 63.5 | |
| Full-ContextCategory=Baseline, Backbone=Qwen3-8B2025.10 | 72.5 | 64.2 | |
| Context-DPOCategory=Training-Based, Backbone=Qwen3-8B2025.10 | 72.3 | 63.9 | |
| CADCategory=Decoding-Based, Backbone=Qwen3-8B2025.10 | 71 | 62 | |
| FaithfulRAGCategory=Prompt-Based, Backbone=LLaMA-3.1-8B-Instruct2025.10 | 70.4 | 65.2 | |
| COIECDCategory=Decoding-Based, Backbone=LLaMA-3.1-8B-Instruct2025.10 | 67.1 | 60.1 | |
| Context-DPOCategory=Training-Based, Backbone=Mistral-7B-v0.32025.10 | 66.4 | 52.7 | |
| COIECDCategory=Decoding-Based, Backbone=Mistral-7B-v0.32025.10 | 58.9 | 45.1 | |
| Full-ContextCategory=Baseline, Backbone=LLaMA-3.1-8B-Instruct2025.10 | 58.5 | 49 | |
| Full-ContextCategory=Baseline, Backbone=Mistral-7B-v0.32025.10 | 58.5 | 44 | |
| AdaCADCategory=Decoding-Based, Backbone=Mistral-7B-v0.32025.10 | 58.2 | 46.3 | |
| AdaCADCategory=Decoding-Based, Backbone=LLaMA-3.1-8B-Instruct2025.10 | 55.7 | 48.1 | |
| CADCategory=Decoding-Based, Backbone=Mistral-7B-v0.32025.10 | 53.5 | 36.9 | |
| CADCategory=Decoding-Based, Backbone=LLaMA-3.1-8B-Instruct2025.10 | 48.3 | 38.1 | |
| No-ContextCategory=Baseline, Backbone=Qwen3-8B2025.10 | 7.8 | 1.4 | |
| No-ContextCategory=Baseline, Backbone=LLaMA-3.1-8B-Instruct2025.10 | 6.1 | 1.3 | |
| No-ContextCategory=Baseline, Backbone=Mistral-7B-v0.32025.10 | 6.1 | 1 |