List your company

Uncertainty Estimation on MMLU AutoGen (test)

0.7315AUROCMATU
0.4926120.5546310.616650.678669Apr 9, 2026

Evaluation results

MethodLinks
MATUBase LLM=Qwen2.5-7B2026.040.73150.8833
SAUP-MultipleBase LLM=Qwen2.5-7B2026.040.71930.8589
Eigv(Agre)-finalBase LLM=Qwen2.5-7B2026.040.71050.8617
Eigv(Agre)-WholeBase LLM=Qwen2.5-7B2026.040.68670.8516
P(true)Base LLM=Qwen2.5-7B2026.040.65560.8363
SAUP-SingleBase LLM=Qwen2.5-7B2026.040.64840.8552
MATUBase LLM=GPT-4o2026.040.62770.5841
Eigv(Agre)-finalBase LLM=GPT-4o2026.040.59810.5649
MATUBase LLM=Llama3.1-8B2026.040.59540.4745
P(true)Base LLM=GPT-4o2026.040.58020.5528
P(true)Base LLM=Llama3.1-8B2026.040.57750.4368
Eigv(Agre)-WholeBase LLM=GPT-4o2026.040.57590.5438
Eigv(Agre)-finalBase LLM=Llama3.1-8B2026.040.55210.4288
Eigv(Agre)-WholeBase LLM=Llama3.1-8B2026.040.53160.3762
SAUP-SingleBase LLM=Llama3.1-8B2026.040.51380.3031
SAUP-MultipleBase LLM=Llama3.1-8B2026.040.50180.2973