List your company

Mathematical Reasoning on MATH (AUROC, AUARC)

0.7121AUROCMATU
0.6178760.6423380.66680.691262Apr 9, 2026

Evaluation results

MethodLinks
MATULLM=Llama3.1-8B2026.040.71210.6518
MATULLM=Qwen2.5-7B2026.040.69660.8585
MATULLM=GPT-4o2026.040.66750.6439
P(true)LLM=Qwen2.5-7B2026.040.66020.8291
P(true)LLM=Llama3.1-8B2026.040.65810.6225
Eigv(Agre)-finalLLM=Qwen2.5-7B2026.040.65060.8212
SAUP-MultipleLLM=Llama3.1-8B2026.040.64770.6065
SAUP-MultipleLLM=Qwen2.5-7B2026.040.63960.8119
Eigv(Agre)-finalLLM=GPT-4o2026.040.63590.6133
Eigv(Agre)-finalLLM=Llama3.1-8B2026.040.6340.6059
SAUP-SingleLLM=Llama3.1-8B2026.040.63390.6008
Eigv(Agre)-WholeLLM=Qwen2.5-7B2026.040.63140.8081
Eigv(Agre)-WholeLLM=GPT-4o2026.040.63080.6115
SAUP-SingleLLM=Qwen2.5-7B2026.040.62610.7982
P(true)LLM=GPT-4o2026.040.62260.607
Eigv(Agre)-WholeLLM=Llama3.1-8B2026.040.62150.5953