Jaehyeok Lee

New framework evaluates LLM cultural value alignment

Researchers have developed a new framework called DOVE to evaluate how well large language models align with cultural values. Unlike previous methods that used multiple-choice questions, DOVE directly compares the distr…