GraphWiz: An Instruction-Following Language Model for Graph Problems

TL;DR

GraphWiz is a specialized language model trained with a new dataset to solve diverse graph problems with explicit reasoning, outperforming GPT-4 and providing insights into training data effects and transferability.

Contribution

We introduce GraphInstruct, a comprehensive instruction dataset, and develop GraphWiz, an open-source model capable of explicit graph problem reasoning, enhanced with DPO for improved accuracy.

Findings

GraphWiz achieves 65% accuracy across nine graph tasks.

GraphWiz surpasses GPT-4's average accuracy of 43.8%.

Training data volume impacts model performance and overfitting.

Abstract

Large language models (LLMs) have achieved impressive success across several fields, but their proficiency in understanding and resolving complex graph problems is less explored. To bridge this gap, we introduce GraphInstruct, a novel and comprehensive instruction-tuning dataset designed to equip language models with the ability to tackle a broad spectrum of graph problems using explicit reasoning paths. Utilizing GraphInstruct, we build GraphWiz, an open-source language model capable of resolving various graph problem types while generating clear reasoning processes. To enhance the model's capability and reliability, we incorporate the Direct Preference Optimization (DPO) framework into the graph problem-solving context. The enhanced model, GraphWiz-DPO, achieves an average accuracy of 65% across nine tasks with different complexity levels, surpassing GPT-4 which has an average accuracy of 43.8%. Moreover, our research delves into the delicate balance between training data volume and model performance, highlighting the potential for overfitting with increased data. We also explore the transferability of the model's reasoning ability across different graph tasks, indicating the model's adaptability and practical application potential. Our investigation offers a new blueprint and valuable insights for developing LLMs specialized in graph reasoning and problem-solving.