Knowledge Conflict Resolution on Conflicts
65.4Exact Match (EM)SDA
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SDAModel=Qwen2.5-7B2026.06 | 65.4 | — | |
| MADARAModel=Qwen2.5-7B, RSC=Stochastic, ρ̂1=0.46, Strategy=SDA2026.06 | 65.4 | 5.5 | |
| 3-AgentModel=Qwen2.5-7B2026.06 | 65 | — | |
| 3-AgentModel=Gemma-2-9B2026.06 | 64.6 | — | |
| CoTModel=Gemma-2-9B2026.06 | 63.3 | — | |
| SDAModel=Gemma-2-9B2026.06 | 63.3 | — | |
| MADARAModel=Gemma-2-9B, RSC=Stochastic, ρ̂1=0.79, Strategy=SDA2026.06 | 63.3 | 2.5 | |
| SDAModel=Qwen3-8B2026.06 | 63 | — | |
| MADARAModel=Qwen3-8B, RSC=Stochastic, ρ̂1=0.39, Strategy=SDA2026.06 | 63 | 2.2 | |
| 3-AgentModel=Qwen3-8B2026.06 | 62.4 | — | |
| CoTModel=Qwen3-8B2026.06 | 62.4 | — | |
| CoTModel=Qwen2.5-7B2026.06 | 62.4 | — | |
| NFModel=Qwen3-8B2026.06 | 60.8 | — | |
| NFModel=Gemma-2-9B2026.06 | 60.8 | — | |
| NFModel=Qwen2.5-7B2026.06 | 59.9 | — | |
| MADARA (Average)2026.06 | 57.1 | 14.4 | |
| MADARAModel=Llama-3.1-8B, RSC=Quality-Ordered, ρ̂1=0.76, Strategy=PDE2026.06 | 50.2 | 36.3 | |
| AverageMethod_Group=3-Agent2026.06 | 46.9 | — | |
| AverageMethod_Group=SDA2026.06 | 46.3 | — | |
| AverageMethod_Group=CoT2026.06 | 45.5 | — | |
| MADARAModel=Mistral-7B-v0.3, RSC=Quality-Ordered, ρ̂1=0.35, Strategy=PDE2026.06 | 43.5 | 25.4 | |
| AverageMethod_Group=NF2026.06 | 42.7 | — | |
| 3-AgentModel=Llama-3.1-8B2026.06 | 24.1 | — | |
| CoTModel=Mistral-7B-v0.32026.06 | 22.8 | — | |
| SDAModel=Mistral-7B-v0.32026.06 | 20.3 | — | |
| SDAModel=Llama-3.1-8B2026.06 | 19.4 | — | |
| 3-AgentModel=Mistral-7B-v0.32026.06 | 18.6 | — | |
| NFModel=Mistral-7B-v0.32026.06 | 18.1 | — | |
| CoTModel=Llama-3.1-8B2026.06 | 16.5 | — | |
| NFModel=Llama-3.1-8B2026.06 | 13.9 | — |