Fact-checking on XSum
56.4Balanced AccuracyANCHOR
Evaluation Results
| Method | Links | |
|---|---|---|
| ANCHORBackbone=Qwen2.5-72B, Direct Decision=false2026.05 | 56.4 | |
| VanillaBackbone=Qwen2.5-72B, Direct Decision=false2026.05 | 55.7 | |
| VanillaBackbone=DeepSeek-V3-671B, Direct Decision=false2026.05 | 53.7 | |
| CoTBackbone=Qwen2.5-72B, Direct Decision=false2026.05 | 53.1 | |
| CoTBackbone=DeepSeek-V3-671B, Direct Decision=false2026.05 | 52.7 | |
| CoTBackbone=DeepSeek-V3-671B, Direct Decision=true2026.05 | 51.7 | |
| BIRDBackbone=Qwen2.5-72B, Direct Decision=false2026.05 | 51.1 | |
| CoTBackbone=Qwen2.5-72B, Direct Decision=true2026.05 | 48.4 | |
| VanillaBackbone=DeepSeek-V3-671B, Direct Decision=true2026.05 | 48.3 | |
| VanillaBackbone=Qwen2.5-72B, Direct Decision=true2026.05 | 47.3 |