MaritimeBench

MaritimeBench 致力于构建一套科学、公平且严谨的航运知识评估体系。基于行业权威标准，我们持续维护并更新高质量的航运数据集——其中包含1,888道客观选择题（MCQ格式），以全面、多维度地量化模型在航运各领域的能力表现。

MaritimeBench is dedicated to establishing a scientific, fair and rigorous maritime knowledge assessment system. Based on authoritative industry standards, we continuously maintain and update high-quality maritime datasets, which comprise 1,888 multiple-choice questions (MCQ format) to comprehensively and multi-dimensionally quantify the performance of models across all maritime domains.

搜集汇总

背景与挑战

背景概述

MaritimeBench是一个航运业专业知识评估数据集，包含1888道选择题，覆盖导航、轮机工程等多个核心领域，旨在提升AI模型的专业能力并支持行业考试和培训。数据集具有权威性、精确分类和全面覆盖等特点。