List your company

Scientific Reasoning on GPQA (test)

11.44Error Rate (%)PAC
11.401611.660811.9212.1792Jan 30, 2026

Evaluation results

MethodLinks
PACScore type=Verbalized score, LLM Backbone=Llama-based, Alpha=0.05, Epsilon=0.22026.0111.44042.52
PACScore type=Router-based score, LLM Backbone=Llama-based, Alpha=0.05, Epsilon=0.22026.0111.89039.67
G-PACScore type=Router-based score, LLM Backbone=Llama-based, Alpha=0.05, Epsilon=0.22026.0111.9039.88
G-PACScore type=Verbalized score, LLM Backbone=Llama-based, Alpha=0.05, Epsilon=0.22026.0111.91036.87
PACScore type=Logits-based score, LLM Backbone=Llama-based, Alpha=0.05, Epsilon=0.22026.0112.387.0318.13
G-PACScore type=Logits-based score, LLM Backbone=Llama-based, Alpha=0.05, Epsilon=0.22026.0112.4023.47