Can Language Models Learn to Skip Steps?

TL;DR

This paper investigates how language models can learn to skip reasoning steps, improving efficiency and generalization, by introducing a framework for training models to generate shorter, accurate reasoning paths.

Contribution

It introduces a novel framework for training language models to learn step-skipping in reasoning, enhancing efficiency and out-of-domain generalization.

Findings

Models can develop step-skipping ability through guided training.

Fine-tuned models resolve tasks more efficiently without losing accuracy.

Models show improved generalization in out-of-domain scenarios.

Abstract

Trained on vast corpora of human language, language models demonstrate emergent human-like reasoning abilities. Yet they are still far from true intelligence, which opens up intriguing opportunities to explore the parallels of humans and model behaviors. In this work, we study the ability to skip steps in reasoning - a hallmark of human expertise developed through practice. Unlike humans, who may skip steps to enhance efficiency or to reduce cognitive load, models do not inherently possess such motivations to minimize reasoning steps. To address this, we introduce a controlled framework that stimulates step-skipping behavior by iteratively refining models to generate shorter and accurate reasoning paths. Empirical results indicate that models can develop the step skipping ability under our guidance. Moreover, after fine-tuning on expanded datasets that include both complete and skipped reasoning sequences, the models can not only resolve tasks with increased efficiency without sacrificing accuracy, but also exhibit comparable and even enhanced generalization capabilities in out-of-domain scenarios. Our work presents the first exploration into human-like step-skipping ability and provides fresh perspectives on how such cognitive abilities can benefit AI models.