Mengxi Guo

Cross-Cultural Scenario Benchmark: Evaluating LLMs’ Cross-Cultural Understanding

Cross-cultural reasoning and alignment have been identified as key weaknesses of large language models (LLMs), but the architectural or cognitive features underlying these failures have not been adequately examined. In addition, previous studies rely almost exclusively on dataset...