Faithfulness Evaluation on WorldCapital Conflicting short-context
127.9AOPCcompATTN
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ATTNModel=Qwen2.5-7B2025.10 | 127.9 | 153.76 | |
| FAModel=Qwen2.5-7B2025.10 | 122.7 | 150.61 | |
| MechLightModel=Qwen2.5-7B2025.10 | 119.3 | 151.61 | |
| IGModel=Qwen2.5-7B2025.10 | 118 | 149.33 | |
| IGModel=Pythia-6.9B2025.10 | 114.5 | 147.82 | |
| FAModel=Pythia-6.9B2025.10 | 113.6 | 147.09 | |
| MechLightModel=Pythia-6.9B2025.10 | 104.2 | 151.9 | |
| ATTNModel=Pythia-6.9B2025.10 | 96.4 | 151.13 |