Explainable Cultural Bias Auditing Framework for Text-to-Image Generative Models Across Multilingual Contexts. JAAIR 2026, 1 (1).