Toward Culturally Aligned LLMs through Ontology-Guided Multi-Agent Reasoning

Wonduk Seo, Wonseok Choi, Junseo Koh +7

Large Language Models (LLMs) increasingly support culturally sensitive decision making, yet often exhibit misalignment due to skewed pretraining data and the absence of structured…