Evaluating Large Language Models with Tests of Spanish as a Foreign Language: Pass or Fail?
This addresses the gap in evaluating LLMs for non-English languages, specifically Spanish, for foreign language learners, though it is incremental as it applies existing methods to new data.
The paper evaluated state-of-the-art LLMs on the TELEIA benchmark, which mimics Spanish exams for foreign students, finding that while LLMs perform well in understanding Spanish, they fall short of native speaker grammatical competence.
Large Language Models (LLMs) have been profusely evaluated on their ability to answer questions on many topics and their performance on different natural language understanding tasks. Those tests are usually conducted in English, but most LLM users are not native English speakers. Therefore, it is of interest to analyze how LLMs understand other languages at different levels: from paragraphs to morphems. In this paper, we evaluate the performance of state-of-the-art LLMs in TELEIA, a recently released benchmark with similar questions to those of Spanish exams for foreign students, covering topics such as reading comprehension, word formation, meaning and compositional semantics, and grammar. The results show that LLMs perform well at understanding Spanish but are still far from achieving the level of a native speaker in terms of grammatical competence.