List your company

Cultural Reasoning on CulturalBench-Hard (CB-H) (test)

46.98AccuracyC-Mining
25.857631.341336.82542.3087Apr 17, 2026

Evaluation results

MethodLinks
C-MiningBackbone=Qwen3-32B, Instruction Tuning Method=C-Mining2026.0446.98
Qwen3-32BBackbone=Qwen3-32B, Instruction Tuning Method=Standard2026.0444.62
C-MiningBackbone=Qwen2.5-7B, Instruction Tuning Method=C-Mining2026.0440.78
CultureBankBackbone=Qwen2.5-7B, Instruction Tuning Method=CultureBank2026.0440.13
CultureLLMBackbone=Qwen2.5-7B, Instruction Tuning Method=CultureLLM2026.0438.99
Llama3.1-8BBackbone=Llama3.1-8B, Instruction Tuning Method=Standard2026.0437.44
Qwen2.5-7BBackbone=Qwen2.5-7B, Instruction Tuning Method=Standard2026.0434.75
GLM4-9BBackbone=GLM4-9B, Instruction Tuning Method=Standard2026.0434.1
CultureSPABackbone=Qwen2.5-7B, Instruction Tuning Method=CultureSPA2026.0430.18
Ministral3-8BBackbone=Ministral3-8B, Instruction Tuning Method=Standard2026.0426.67