LaRS: Latent Reasoning Skills for Chain-of-Thought Reasoning

TL;DR

LaRS introduces an unsupervised method to identify and select reasoning skills in large language models, improving reasoning performance and efficiency without manual prompt design or auxiliary inference.

Contribution

It proposes a novel unsupervised approach to model and select latent reasoning skills, enhancing chain-of-thought prompting efficiency and robustness.

Findings

Outperforms state-of-the-art skill-based selection methods

Processes example banks four times faster

Reduces LLM inferences during selection by half

Abstract

Chain-of-thought (CoT) prompting is a popular in-context learning (ICL) approach for large language models (LLMs), especially when tackling complex reasoning tasks. Traditional ICL approaches construct prompts using examples that contain questions similar to the input question. However, CoT prompting, which includes crucial intermediate reasoning steps (rationales) within its examples, necessitates selecting examples based on these rationales rather than the questions themselves. Existing methods require human experts or pre-trained LLMs to describe the skill, a high-level abstraction of rationales, to guide the selection. These methods, however, are often costly and difficult to scale. Instead, this paper introduces a new approach named Latent Reasoning Skills (LaRS) that employs unsupervised learning to create a latent space representation of rationales, with a latent variable called a reasoning skill. Concurrently, LaRS learns a reasoning policy to determine the required reasoning skill for a given question. Then the ICL examples are selected by aligning the reasoning skills between past examples and the question. This approach is theoretically grounded and compute-efficient, eliminating the need for auxiliary LLM inference or manual prompt design. Empirical results demonstrate that LaRS consistently outperforms SOTA skill-based selection methods, processing example banks four times faster, reducing LLM inferences during the selection stage by half, and showing greater robustness to sub-optimal example banks.