SOURCE-LINKED INTELLIGENCE

Cultural Misalignment in Large Language Models: Detection, Measurement, and Mitigation Through Targeted Fine-Tuning

arXiv · AI, language, vision and robotics · article · Sep 3, 2026 · UTC

We evaluate three open-weight LLMs (Gemma3-12B from the USA, Bielik-11B-v3 from Poland, and Qwen3-4B from China) against World Values Survey Wave 7 data for 63 demographic personas across three countries, using normalized Wasserstein distance to quantify distributional misalignment. Contrary to expectations, no model favors its home country: the Chinese-built Qwen3-4B performs worst on its own Chinese population (W1 = 0.436, the highest misalignment in the entire model x country matrix). Targeted LoRA fine-tuning on the five worst-case personas, requiring fewer than 1,200 training pairs and un

Read original source ↗ Open in workspace

recordType

paper

region

Global