Building a domain-specific large language model for Chinese ancient books: a case study on epitaph modelling
DOI:
https://doi.org/10.47989/ir31361795Keywords:
Generative Artificial Intelligence, Domain-Specific, Large Language Model, Ancient Books, EpitaphAbstract
Introduction. With the rapid advancement of generative artificial intelligence technology, large language models (LLMs) are expected to play an important role in improving access to and the use of library collections. This study constructs and develops a generative conversational large language model tailored specifically to epitaphs, overcoming traditional research paths and systems of digitisation.
Method. Based on the Chinese epitaph database built by the Zhejiang University Library and some commercial epitaph databases, an epitaph generative dialogue large language model was constructed using Doubao as the base model through pre-training, supervised fine-tuning, and other full process trainings.
Analysis. This study used human evaluation, in which experts, researchers, or general users assessed the quality of model-generated outputs. The evaluation included the epitaphs of Cui Yi, Liu Yi, and Lihua in the Tang dynasty.
Results. The model successfully performed intelligent interpretation of epitaphs, automatic information extraction, and other knowledge-based services, thereby providing new ideas for constructing an ancient literature corpus from excavated materials represented by epitaphs.
Conclusion. A large-scale domain-specific foundational model tailored for library applications was successfully developed. The study offers innovative approaches for processing historical texts and suggests new directions for AI integration in library knowledge systems.
References
Brzustowicz, R. (2023). From ChatGPT to CatGPT: The implications of artificial intelligence on library cataloging. Information Technology and Libraries, 42(3). https://doi.org/10.5860/ital.v42i3.16295
Central Committee of the Communist Party of China, & General Office of the State Council. (2022). Opinions on promoting ancient books work in the new era. https://www.gov.cn/gongbao/content/2022/content_5687500.htm (Archived at https://web.archive.org/web/20250901180020/https://www.gov.cn/gongbao/content/2022/content_5687500.htm )
Chen H., Chen, H., Han, K., Zhu, G., Zhao, Y., & Du, Y. (2024). Customization of vertical domain large models: Theoretical basis and key technologies. Journal of Data Acquisition and Processing, 39(3), 524–546. https://doi.org/10.16337/j.1004-9037.2024.03.003
Chen, L. (2022). Issues in the digitization of ancient books and the construction of classical knowledge bases from the perspective of digital humanities. Journal of Library Science in China, 48(2), 36–46. https://doi.org/10.13530/j.cnki.jlis.2022013
Guo, Y., Kou, X., Feng, S., & Li, S. (2024). Empowering library reference services with large language models: Logic, scenarios, and systems. Library Tribune, 1–10.
Hu, H. F., & Tang, Y. (2013). Catalogue of epitaph rubbings from various dynasties in Peking University Library. Shanghai Chinese Classics Publishing House.
Huang, Q., Tao, M., Zhang, C., An, Z., Jiang, C., Chen, Z., Wu, Z., & Feng, Y. (2023). Lawyer LLaMA technical report. arXiv. https://doi.org/10.48550/arXiv.2305.15062
Lappalainen, Y., & Narayanan, N. (2023). Aisha: A custom AI library chatbot using the ChatGPT API. Journal of Web Librarianship, 17, 37–58. https://doi.org/10.1080/19322909.2023.2221477
Li, H. Y. (2019). Research on restoration methods of digital epitaph rubbings [Master’s thesis, China West Normal University].
Li, S., Hu, R., & Wang, L. (2024). Construction and application research of ancient Chinese large language models. Language Strategy Research, 9(5), 22–33.
Luo, Y., Yang, K., Hong, M., Liu, X. Y., & Nie, Z. (2023). MolFM: A multimodal molecular foundation model. arXiv. https://doi.org/10.48550/arXiv.2307.09484
Luo, Y., Zhang, J., Fan, S., Yang, K., Hong, M., Wu, Y., Qiao, M., & Nie, Z. (2024). BiomedGPT: An open multimodal large language model for biomedicine. IEEE Journal of Biomedical and Health Informatics. https://doi.org/10.1109/JBHI.2024.3505955
Rong, L. (2024). Research on prompt learning for automatic classification of books using large language models. Library Science Research, (1), 86–103. https://doi.org/10.15941/j.cnki.issn1001-0424.2024.01.012
Song, Z., Yan, B., Liu, Y., Fang, M., Li, M., Yan, R., & Chen, X. (2025). Injecting domain-specific knowledge into large language models: A comprehensive survey. In Findings of the Association for Computational Linguistics: EMNLP 2025 (pp. 25297–25311). Association for Computational Linguistics. https://doi.org/10.18653/v1/2025.findings-emnlp.1379
Verma, M. (2023). Novel study on AI-based chatbot (ChatGPT) impacts on the traditional library management. International Journal of Trend in Scientific Research and Development, 7(1), 961–964. https://bit.ly/4xdGtP4 (Archived by the Internet Archive at https://bit.ly/3UZWgni )
Wang, D. (2023). XunziALLM [Large language model]. GitHub. https://github.com/Xunzi-LLM-of-Chinese-classics/XunziALLM
Wu, N., Liu, C., Liu, J. F., & Wang, D. B. (2024). AIGC-driven automatic summarization of ancient texts: From natural language understanding to generation. Library Tribune, 44(9), 111–123.
Wu, X., Yan, D., & Cai, Y. (2024). Innovative practice of AI literacy education based on the ARCS model. Library Journal, 1–12. https://doi.org/10.13663/j.cnki.lj.2024.11.006
Wang, X. (n.d.). Mu ming ju li [Examples of epitaph compositions]. In Complete library of the four treasuries (Wen-Yuan-Ge ed., Vol. 1482). Shanghai Classics Publishing House. (Original work published during the Ming dynasty).
Xiao, T. (2025). Wenxuan [Anthology of literature] (Facsimile reprint). Chongwen Publishing House. (Original work published ca. 520–530 C.E.)
Xu, Q. K., Wang, D. B., Liu, Y. T., & Huang, S. Q. (2024). Research on the construction of ancient books knowledge base for knowledge services. Information Science, 42(5), 149–158. https://doi.org/10.13833/j.issn.1007-7634.2024.05.016
Zhang, H., Chen, J., Jiang, F., Yu, F., Chen, Z., Li, J., Chen, G., Wu, X., Zhang, Z., Xiao, Q., Wan, X., Wang, B., & Li, H. (2023). HuatuoGPT, towards taming language model to be a doctor. arXiv. https://doi.org/10.18653/v1/2023.findings-emnlp.725
Zhang, J. D., Yang, S. H., Liu, J. F., & Huang, Q. (2024). Empowering traditional Chinese medicine ancient texts with AIGC: Construction of the Huang-Di large model. Library Tribune, 44(10), 103–112.
Zhang, X., Tian, C., Yang, X., Chen, L., Li, Z., & Petzold, L. R. (2023). AlpaCare: Instruction-tuned large language models for medical application. arXiv. https://doi.org/10.48550/arXiv.2310.14558
Zhao, S., Zhang, J., & Nie, Z. (2023). Large-scale cell representation learning via divide-and-conquer contrastive learning. arXiv. https://doi.org/10.48550/arXiv.2306.04371
Zhao, S. Q., Zhou, Y., Ren, Y., Chen, Z., Jia, C., Fang, Z., Long, Z., Liu, S., & Lan, M. (2025). Fùxì: A benchmark for evaluating language models on ancient Chinese text understanding and generation. arXiv. https://doi.org/10.48550/arxiv.2503.15837
Zhao, Y. (2018). Preliminary discussion on the digital processing of Tang Dynasty epitaphs. Library Tribune, 38(6), 49–55.
Zhao, Z. X., Hu, D., Liu, C., Shen, S., & Wang, D. B. (2024). Performance evaluation of Chinese general large models in humanities and social sciences. Library and Information Service, 68(13), 132–143. https://doi.org/10.13266/j.issn.0252-3116.2024.13.012
Zuo, L., Zhao, Z., & Wang, D. (2024). Automatic classification of "Siku Quanshu" based on large language models. Journal of Information Resources Management, 14(5), 22–35. https://doi.org/10.13365/j.jirm.2024.05.023
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Minjie Yu, Haoyue jin, Shiliang Xue
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
