Enhancing Reasoning Capabilities of Small Language Models with Blueprints and Prompt Template Search

TL;DR

This paper introduces a framework that enhances small language models' reasoning abilities by using LLM-generated blueprints and prompt template search, improving performance across multiple tasks without increasing model size.

Contribution

The proposed framework leverages high-level blueprints and prompt template search to significantly improve small language models' reasoning capabilities without additional training.

Findings

Improved reasoning performance on math, coding, and logic tasks.

Enhanced robustness to prompt variations.

No increase in model size or training required.

Abstract

Small language models (SLMs) offer promising and efficient alternatives to large language models (LLMs). However, SLMs' limited capacity restricts their reasoning capabilities and makes them sensitive to prompt variations. To address these challenges, we propose a novel framework that enhances SLM reasoning capabilities through LLM generated blueprints. The blueprints provide structured, high-level reasoning guides that help SLMs systematically tackle related problems. Furthermore, our framework integrates a prompt template search mechanism to mitigate the SLMs' sensitivity to prompt variations. Our framework demonstrates improved SLM performance across various tasks, including math (GSM8K), coding (MBPP), and logic reasoning (BBH). Our approach improves the reasoning capabilities of SLMs without increasing model size or requiring additional training, offering a lightweight and deployment-friendly solution for on-device or resource-constrained environments.