List your company

Cultural commonsense reasoning on CultureAtlas (All Culture)

95.8PrecisionGPT-4
62.31271.00679.788.394Jan 7, 2026

Evaluation results

MethodLinks
GPT-42026.0195.890.693.1
CALM2026.0193.687.789.1
LLaMA-2Parameters=7B2026.0184.242.156.1
VicunaParameters=7B2026.0179.656.866.3
VicunaParameters=13B2026.0167.481.273.7
LLaMA-2Parameters=13B2026.0163.677.169.7