Boosting Logical Reasoning in Large Language Models through a New Framework: The Graph of Thought

TL;DR

The paper introduces the Graph of Thoughts (GoT), a novel prompting framework that significantly enhances large language models' multi-step logical reasoning abilities across various complex tasks.

Contribution

It presents the Graph of Thoughts (GoT), a new prompting technique that outperforms existing methods like Tree of Thought (ToT) in logical reasoning tasks.

Findings

GoT achieves up to 89.7% accuracy on complex reasoning tasks.

GoT outperforms ToT with an average accuracy increase of around 20%.

The method demonstrates substantial improvements across diverse logical challenges.

Abstract

Recent advancements in large-scale models, such as GPT-4, have showcased remarkable capabilities in addressing standard queries. However, when facing complex problems that require multi-step logical reasoning, their accuracy dramatically decreases. Current research has explored the realm of \textit{prompting engineering} to bolster the inferential capacities of these models. Our paper unveils a pioneering prompting technique, dubbed \textit{Graph of Thoughts (GoT)}. Through testing on a trio of escalating challenges: the 24-point game, resolution of high-degree polynomial equations, and derivation of formulas for recursive sequences, our method outperformed GPT-4, achieving accuracy improvements of $89.7\%$, $86\%$, and $56\%$ for each respective task. Moreover, when juxtaposed with the state-of-the-art (SOTA) prompting method, \textit{Tree of Thought (ToT)}, our approach registered an average accuracy boost of $23\%$, $24\%$, and $15\%$.