Question Answering on 2WikiMHQA (test)
45.8EMVEROIC
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| VEROICInference Paths=LLaMA-3.1-8B -> LLaMA-3.1-70B2026.04 | 45.8 | 62.6 | |
| VEROICInference Paths=Qwen-2.5-7B -> Qwen-2.5-32B2026.04 | 45.8 | 62.3 | |
| RISK PREDICTORInference Paths=LLaMA-3.1-8B -> LLaMA-3.1-70B2026.04 | 45.3 | 62 | |
| RISK PREDICTORInference Paths=Qwen-2.5-7B -> Qwen-2.5-32B2026.04 | 45.1 | 61.8 | |
| CONFIDENCE-THRESHOLDInference Paths=LLaMA-3.1-8B -> LLaMA-3.1-70B2026.04 | 44.8 | 61.6 | |
| OUTPUT-TRIGGERInference Paths=LLaMA-3.1-8B -> LLaMA-3.1-70B2026.04 | 44.7 | 61.4 | |
| CONFIDENCE-THRESHOLDInference Paths=Qwen-2.5-7B -> Qwen-2.5-32B2026.04 | 44.7 | 61 | |
| OUTPUT-TRIGGERInference Paths=Qwen-2.5-7B -> Qwen-2.5-32B2026.04 | 44.5 | 60.8 | |
| PERIODICInference Paths=LLaMA-3.1-8B -> LLaMA-3.1-70B2026.04 | 43.9 | 60.3 | |
| RANDOMInference Paths=LLaMA-3.1-8B -> LLaMA-3.1-70B2026.04 | 43.4 | 59.7 | |
| PERIODICInference Paths=Qwen-2.5-7B -> Qwen-2.5-32B2026.04 | 43.1 | 59.5 | |
| RANDOMInference Paths=Qwen-2.5-7B -> Qwen-2.5-32B2026.04 | 42.8 | 59.2 | |
| FIXEDInference Paths=LLaMA-3.1-8B -> LLaMA-3.1-70B2026.04 | 41.2 | 57.6 | |
| FIXEDInference Paths=Qwen-2.5-7B -> Qwen-2.5-32B2026.04 | 40.5 | 56.9 |