BEnQA

BEnQA数据集是由韩国科学技术院创建的，包含5161条平行于孟加拉语和英语的中学和高中水平考试题目。该数据集覆盖了多个科学学科，包括事实性、应用性和基于推理的问题。数据集的创建过程涉及从印刷材料中收集考试题目，并通过熟练的双语打字员进行数字化。BEnQA数据集主要用于评估大型语言模型在低资源语言中的表现，特别是在教育和科技领域的应用，旨在解决语言模型在非英语语言中的性能差异问题。

The BEnQA dataset was developed by the Korea Advanced Institute of Science and Technology (KAIST), consisting of 5,161 parallel exam questions in Bengali and English for middle and high school levels. This dataset covers multiple scientific disciplines, including factual, applied, and reasoning-based questions. The dataset construction process involved collecting exam questions from printed materials and digitizing them via skilled bilingual typists. Primarily intended to evaluate the performance of large language models (LLMs) in low-resource languages, especially for educational and technological applications, the BEnQA dataset aims to address the performance disparities of language models across non-English languages.