InstructGraph: Boosting Large Language Models via Graph-centric Instruction Tuning and Preference Alignment

TL;DR

InstructGraph is a novel framework that enhances large language models' ability to perform graph reasoning and generation through instruction tuning, a universal graph data format, and preference alignment, significantly outperforming existing models.

Contribution

The paper introduces a unified graph verbalizer, a dedicated instruction tuning process, and a preference alignment strategy to improve LLMs' graph reasoning capabilities.

Findings

InstructGraph outperforms GPT-4 and LLaMA2 by over 13% and 38%.

The universal code-like format simplifies graph data representation.

Preference alignment reduces hallucinations in graph tasks.

Abstract

Do current large language models (LLMs) better solve graph reasoning and generation tasks with parameter updates? In this paper, we propose InstructGraph, a framework that empowers LLMs with the abilities of graph reasoning and generation by instruction tuning and preference alignment. Specifically, we first propose a structured format verbalizer to unify all graph data into a universal code-like format, which can simply represent the graph without any external graph-specific encoders. Furthermore, a graph instruction tuning stage is introduced to guide LLMs in solving graph reasoning and generation tasks. Finally, we identify potential hallucination problems in graph tasks and sample negative instances for preference alignment, the target of which is to enhance the output's reliability of the model. Extensive experiments across multiple graph-centric tasks exhibit that InstructGraph can achieve the best performance and outperform GPT-4 and LLaMA2 by more than 13\% and 38\%, respectively.