Abstraction-of-Thought Makes Language Models Better Reasoners

TL;DR

This paper introduces Abstraction-of-Thought (AoT), a novel reasoning format that enhances language models' ability to perform abstract reasoning, leading to improved performance on complex tasks compared to traditional methods.

Contribution

The paper proposes AoT, a new structured reasoning format, along with a large finetuning dataset, demonstrating significant performance improvements on challenging reasoning benchmarks.

Findings

AoT outperforms Chain-of-Thought in many reasoning tasks

A new dataset of 348k AoT samples was created for finetuning

Models trained with AoT show improved generalization on Big-Bench Hard

Abstract

Abstract reasoning, the ability to reason from the abstract essence of a problem, serves as a key to generalization in human reasoning. However, eliciting language models to perform reasoning with abstraction remains unexplored. This paper seeks to bridge this gap by introducing a novel structured reasoning format called Abstraction-of-Thought (AoT). The uniqueness of AoT lies in its explicit requirement for varying levels of abstraction within the reasoning process. This approach could elicit language models to first contemplate on the abstract level before incorporating concrete details, which is overlooked by the prevailing step-by-step Chain-of-Thought (CoT) method. To align models with the AoT format, we present AoT Collection, a generic finetuning dataset consisting of 348k high-quality samples with AoT reasoning processes, collected via an automated and scalable pipeline. We finetune a wide range of language models with AoT Collection and conduct extensive evaluations on 23 unseen tasks from the challenging benchmark Big-Bench Hard. Experimental results indicate that models aligned to AoT reasoning format substantially outperform those aligned to CoT in many reasoning tasks.