List your company

Uncertainty Estimation on MATH AutoGen (test)

0.7544AUROCMATU
0.5176960.5791480.64060.702052Apr 9, 2026

Evaluation results

MethodLinks
MATUBase LLM=Llama3.1-8B2026.040.75440.4687
MATUBase LLM=Qwen2.5-7B2026.040.71460.5334
MATUBase LLM=GPT-4o2026.040.65820.622
P(true)Base LLM=Qwen2.5-7B2026.040.65240.5102
Eigv(Agre)-finalBase LLM=Qwen2.5-7B2026.040.63550.4512
SAUP-MultipleBase LLM=Llama3.1-8B2026.040.63340.3933
P(true)Base LLM=Llama3.1-8B2026.040.62710.4571
Eigv(Agre)-WholeBase LLM=Qwen2.5-7B2026.040.61110.4326
P(true)Base LLM=GPT-4o2026.040.60790.5931
SAUP-SingleBase LLM=Llama3.1-8B2026.040.60640.383
Eigv(Agre)-WholeBase LLM=GPT-4o2026.040.60150.5892
Eigv(Agre)-finalBase LLM=Llama3.1-8B2026.040.59120.3802
Eigv(Agre)-finalBase LLM=GPT-4o2026.040.58980.5826
Eigv(Agre)-WholeBase LLM=Llama3.1-8B2026.040.57610.3679
SAUP-MultipleBase LLM=Qwen2.5-7B2026.040.53850.409
SAUP-SingleBase LLM=Qwen2.5-7B2026.040.52680.399