TaiwanVQA: Benchmarking and Enhancing Cultural Understanding in Vision-Language Models | Read Paper on Bytez