<i>Decoding Nested Entities from Classical Chinese with LLMs.</i>

This repository contains the datasets, source codes, and experimental results for the study <i>“Decoding Nested Entities from Classical Chinese with LLMs.”</i><br>The research focuses on extracting and analyzing nested entities from large-scale Chinese local gazetteers. We design a data sampling strategy integrated with the multi-objective optimization algorithm NSGA-II to address entity imbalance in historical low-resource corpora. Two Large Language Models, Qwen 2.5-7B and Xunzi-Qwen2-7B, are fine-tuned with LoRA and few-shot prompt learning, and compared with non-generative baselines (GlobalPointer, Biaffine, TPLinker).