MARK18964758

This dataset contains the complete text of several novel, fully generated with the assistance of open-source large language models (LLMs). It represents an experiment in large-scale literary creation under a human-AI collaborative paradigm: the human author designed the story architecture, character settings, historical research, and narrative pacing, while open-source LLMs carried out text expansion, dialogue generation, scene depiction, and the weaving of multiple narrative threads based on detailed prompts. In addition to individual files, this dataset is also packaged in JSON format ( )for large‑scale training, strictly following the instruction‑tuning data formats of mainstream LLMs…