albertge/synthetic-orqa
收藏官方服务:
资源简介:
合成运筹学问题回答数据集包含946个运筹学问题,分为训练和验证两部分。总样本数为946,其中训练样本901个,是通过GPT-5和Claude Sonnet 4.5生成的合成数据;验证样本45个,是原始种子数据。该数据集的任务是运筹学推理和优化问题识别。
This dataset contains 946 Operations Research Question Answering (ORQA) problems with train/validation splits. There are a total of 946 samples, of which 901 are training samples (synthetic, generated using GPT-5 and Claude Sonnet 4.5), and 45 are validation samples (original seed data). The task is to identify operations research reasoning and optimization problems.
提供机构:
albertge


