Year: 2026
Type: conference-paper
Abstract: Automatic instruction generation offers a lowcost, high-efficiency pathway for fine-tuning large language models (LLMs).However, existing methods struggle in knowledge-intensive domains and complex reasoning tasks due to their dependence on high-quality seed data, limited coverage of single-document knowledge, and repetitive content.To overcome these limitations, this paper presents GCIG, a GraphRAG-based Cross-document Instruction Generation framework.We begin by constructing an enhanced knowledge graph to provide a structural representation of the raw corpus, followed by LLM-driven selection of reliable subgraph-text pairs based on factuality and logical complementarity.Subsequently, we adaptively generate diverse questions through task-aware prompts and context-sensitive retrieval.Finally, we employ Chain-of-Thought reasoning to anchor entity paths and integrate scattered evidence, thereby closing logical gaps and improving answer coherence.Experiments on knowledgeintensive and multi-hop question-answering tasks demonstrate that GCIG outperforms existing methods, producing instruction data with stronger logical consistency and broader knowledge coverage for effective LLM fine-tuning.
Authors: Xiaoliang Xu, Huang Yuan, Junmei Wang, Can Xu
Institution: Hangzhou Dianzi University
Language: English
Topic: Topic Modeling
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
Awards: G2628709859, G5099709406
OpenAlex ID: W7166813354