List your company

Question Answering on PopQA (AUROC, AUARC, ECE)

0.908AUROCSGPU
0.654240.720120.7860.85188Dec 16, 2025

Evaluation results

MethodLinks
SGPULLM Architecture=Llama-3.1-8B2025.120.9080.9020.42
UMPIRELLM Architecture=Llama-3.1-8B2025.120.8710.570.05
Cos EigenscoreLLM Architecture=Llama-3.1-8B2025.120.8530.8460.132
Cov EigenscoreLLM Architecture=Llama-3.1-8B2025.120.7840.7530.114
SELLM Architecture=Llama-3.1-8B2025.120.7390.6180.13
DSELLM Architecture=Llama-3.1-8B2025.120.7340.6230.159
ManualMaternLLM Architecture=Llama-3.1-8B2025.120.7050.6260.656
RegularLLM Architecture=Llama-3.1-8B2025.120.6780.6460.334
ManualHeatLLM Architecture=Llama-3.1-8B2025.120.6640.6460.455