Can Separators Improve Chain-of-Thought Prompting?

TL;DR

This paper introduces COT-SEP, a novel method that adds separators in chain-of-thought prompting to enhance reasoning performance of large language models on complex tasks.

Contribution

The paper proposes a separator-based enhancement to chain-of-thought prompting, improving reasoning accuracy and understanding in LLMs across various models and tasks.

Findings

COT-SEP significantly outperforms vanilla CoT on reasoning tasks.

Separator type and placement affect LLM performance.

Effective separators improve reasoning clarity and accuracy.

Abstract

Chain-of-thought (CoT) prompting is a simple and effective method for improving the reasoning capabilities of Large Language Models (LLMs). The basic idea of CoT is to let LLMs break down their thought processes step-by-step by putting exemplars in the input prompt. However, the densely structured prompt exemplars of CoT may cause the cognitive overload of LLMs. Inspired by human cognition, we introduce COT-SEP, a method that strategically employs separators at the end of each exemplar in CoT prompting. These separators are designed to help the LLMs understand their thought processes better while reasoning. Interestingly, it turns out that COT-SEP significantly improves the LLMs' performances on complex reasoning tasks (e.g., GSM8K, AQuA, CSQA), compared with the vanilla CoT, which does not use separators. We also study the effects of the type and the location of separators tested on multiple LLMs, including GPT-3.5-Turbo, GPT-4, and LLaMA-2 7B.