Evaluating LLMs' Multilingual Capabilities for Bengali: Benchmark Creation and Performance Analysis
This work addresses the problem of underrepresented languages in NLP for Bengali speakers and researchers, though it is incremental as it focuses on benchmark creation and analysis rather than developing new methods.
The researchers tackled the lack of standardized benchmarks for Bengali NLP by creating one and evaluating 10 open-source LLMs, finding consistent performance gaps compared to English, with smaller models and Mistral family showing particular weaknesses, while DeepSeek maintained more stable performance.
Bengali is an underrepresented language in NLP research. However, it remains a challenge due to its unique linguistic structure and computational constraints. In this work, we systematically investigate the challenges that hinder Bengali NLP performance by focusing on the absence of standardized evaluation benchmarks. We then evaluated 10 recent open source Large Language Models (LLMs) in 8 of the translated datasets and performed a comprehensive error analysis to pinpoint their primary failure modes. Our findings reveal consistent performance gaps for Bengali compared to English, particularly for smaller models and specific model families like Mistral. We also identified promising robustness in certain architectures, such as DeepSeek, that maintain more stable performance across languages. Our analysis reveals an inverse relationship between tokenization efficiency and LLM accuracy where models tend to perform worse when inputs are excessively tokenized, whereas more efficient \& concise tokenization results in improved performance. These findings highlight critical areas where current models fall short and underscore the need for improved dataset quality and evaluation methodologies tailored to multilingual contexts. This work will catalyze further research on NLP for underrepresented languages, helping to democratize access to advanced language technologies worldwide. The code and dataset used in this research is publicly available at https://github.com/BengaliAI/bn-llm-benchmark.