MasalBench: A Benchmark for Contextual and Cross-Cultural Understanding of Persian Proverbs in LLMs
MasalBench is a benchmark created to test how well multilingual Large Language Models (LLMs) understand Persian proverbs in context and across cultures. The study found that LLMs excel at understanding Persian proverbs within their language but struggle significantly when asked to find equivalent English proverbs, highlighting limitations in cultural knowledge.