Benchmark RadarAI BENCHMARK PROFILE
CultureForest
CultureForest is a benchmark for cultural norm grounded reasoning, featuring 5,378 examples across 8 domains and 53 countries/regions, with progressive evaluation from multiple-choice to open-ended generation.
- Released
- 2026-06-01
- Readiness
- Paper only
- Primary field
- General AI
Why it matters
The benchmark addresses the evaluation gap between cultural knowledge and its application, offering a verifiable and attributable assessment of reasoning grounded in cultural norms.
Motivation
Existing research largely reduces cultural intelligence in LLMs to a knowledge-level problem, overlooking whether models can effectively utilize their acquired knowledge in realistic scenarios.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.