List your company

Question Answering on 2WikiMHQA (test)

45.8EMVEROIC
40.28841.71943.1544.581Apr 30, 2026

Evaluation results

MethodLinks
VEROICInference Paths=LLaMA-3.1-8B -> LLaMA-3.1-70B2026.0445.862.6
VEROICInference Paths=Qwen-2.5-7B -> Qwen-2.5-32B2026.0445.862.3
RISK PREDICTORInference Paths=LLaMA-3.1-8B -> LLaMA-3.1-70B2026.0445.362
RISK PREDICTORInference Paths=Qwen-2.5-7B -> Qwen-2.5-32B2026.0445.161.8
CONFIDENCE-THRESHOLDInference Paths=LLaMA-3.1-8B -> LLaMA-3.1-70B2026.0444.861.6
OUTPUT-TRIGGERInference Paths=LLaMA-3.1-8B -> LLaMA-3.1-70B2026.0444.761.4
CONFIDENCE-THRESHOLDInference Paths=Qwen-2.5-7B -> Qwen-2.5-32B2026.0444.761
OUTPUT-TRIGGERInference Paths=Qwen-2.5-7B -> Qwen-2.5-32B2026.0444.560.8
PERIODICInference Paths=LLaMA-3.1-8B -> LLaMA-3.1-70B2026.0443.960.3
RANDOMInference Paths=LLaMA-3.1-8B -> LLaMA-3.1-70B2026.0443.459.7
PERIODICInference Paths=Qwen-2.5-7B -> Qwen-2.5-32B2026.0443.159.5
RANDOMInference Paths=Qwen-2.5-7B -> Qwen-2.5-32B2026.0442.859.2
FIXEDInference Paths=LLaMA-3.1-8B -> LLaMA-3.1-70B2026.0441.257.6
FIXEDInference Paths=Qwen-2.5-7B -> Qwen-2.5-32B2026.0440.556.9