2026-10-12T09:20:43Z https://www.scidb.cn/oai

10.57760/sciencedb.j00104.00130 2026-09-10T10:53:50Z

2026-09-10 A dataset of Mongolian reward training Q&A doi:10.57760/sciencedb.j00104.00130 en We collect the existing English reward training corpus, and constructs 11,000 Mongolian reward training corpus through the combination of rule screening, machine translation, manual construction, program correction and manual proofreading. The corpus mainly focuses on the daily dialogue in the open field. The publication of this dataset has important usage value for promoting the development of Mongolian large language models. Minority Languages; Low-Resource Language; Mongolian; Dataset; LLM Luo Jian Sun Yuan Zhao Xiaobing PUBLIC https://creativecommons.org/licenses/by-nc/4.0/ dataset Science Data Bank