batuhanozkose/Rehber-CoT-Science-v1

Rehber-CoT-Science-v1是一个高质量的土耳其语科学领域计算推理数据集，旨在提升土耳其语在科学领域的计算推理能力。数据集包含六个科学领域（物理、化学、生物、数学、工程和科学）的问题，具有两种难度级别（博士和本科），每个问题都附有可执行的Python代码和详细的解题步骤（Chain-of-Thought）。数据集经过三阶段混合验证（数值匹配、正则匹配和LLM判断），确保高质量和准确性。数据集还包含匿名化的源引用和唯一的UUID标识符。

Rehber-CoT-Science-v1 is a high-quality Turkish-language scientific computational reasoning dataset designed to enhance computational reasoning capabilities in Turkish across various scientific domains. The dataset includes questions from six scientific domains (Physics, Chemistry, Biology, Mathematics, Engineering, and Science) with two difficulty levels (PhD and Undergraduate). Each question comes with executable Python code and detailed solution steps (Chain-of-Thought). The dataset undergoes a three-stage hybrid verification process (numeric match, regex match, and LLM judge) to ensure high quality and accuracy. It also includes anonymized source references and unique UUID identifiers.