List your company

Human process distribution matching on Sampling Task

0.15Mean Absolute Cohen's dCentaur
0.10560.40530.7051.0047May 7, 2026

Evaluation results

MethodLinks
CentaurParameter Scale=70B2026.050.15
GPT-52026.050.27
Base QwenParameter Scale=1.5B2026.050.89
Claude Sonnet2026.051.15
Gemini 2.5 Pro2026.051.26