Abstract: [Purpose/Significance] This study aims to apply generative artificial intelligence（GenAI）to uncover hidden knowledge associations in ancient texts and achieve multi-dimensional reorganization and structural regeneration, facilitating their revitalized use in the era of digital intelligence. [Design/Methodology] This study proposed a multi-dimensional knowledge reorganization framework for ancient texts based on generative logic. First, a multi-dimensional knowledge representation model was built using YAML-based markup semantic framework. Second, the large language model ChatGPT-4 Turbo, in conjunction with the SPIRES algorithm, was employed to automatically extract multi-dimensional instances. Finally, by integrating the Neo4j graph database, the deep associations among knowledge elements across different dimensions were visualized through relational graphs, offering readers a viable solution for facilitating efficient and immersive reading experience. [Findings/Conclusion] Taking Shiji: Liezhuan（The Biographies）as an example, the study validates the rationality of the semantic markup framework in multi-dimensional knowledge modeling of ancient texts, thereby achieving the representation of intrinsic knowledge associations and their multi-dimensional reorganization from structured descriptions. [Originality/Value] This study proposes a YAML-based generative guidance mechanism and a multi-dimensional knowledge extraction framework, which expands the application boundaries of generative artificial intelligence in the field of digital humanities and offers a new technical paradigm for the theory of knowledge organization and communication.

Keywords: Ancient classic texts, Large Language Models, Knowledge reorganization, Digital humanities, Knowledge organization