Active Prompting with Chain-of-Thought for Large Language Models

TL;DR

This paper introduces Active-Prompt, a method that uses active learning principles to select the most informative questions for annotation, enhancing chain-of-thought prompting in large language models and achieving state-of-the-art results on reasoning tasks.

Contribution

It proposes an active learning approach to optimize prompt examples for LLMs, improving reasoning performance without relying solely on human-annotated exemplars.

Findings

Achieves state-of-the-art on eight reasoning tasks

Effective question selection improves model performance

Uncertainty metrics guide optimal example annotation

Abstract

The increasing scale of large language models (LLMs) brings emergent abilities to various complex tasks requiring reasoning, such as arithmetic and commonsense reasoning. It is known that the effective design of task-specific prompts is critical for LLMs' ability to produce high-quality answers. In particular, an effective approach for complex question-and-answer tasks is example-based prompting with chain-of-thought (CoT) reasoning, which significantly improves the performance of LLMs. However, current CoT methods rely on a fixed set of human-annotated exemplars, which are not necessarily the most effective examples for different tasks. This paper proposes a new method, Active-Prompt, to adapt LLMs to different tasks with task-specific example prompts (annotated with human-designed CoT reasoning). For this purpose, we propose a solution to the key problem of determining which questions are the most important and helpful ones to annotate from a pool of task-specific queries. By borrowing ideas from the related problem of uncertainty-based active learning, we introduce several metrics to characterize the uncertainty so as to select the most uncertain questions for annotation. Experimental results demonstrate the superiority of our proposed method, achieving state-of-the-art on eight complex reasoning tasks. Further analyses of different uncertainty metrics, pool sizes, zero-shot learning, and accuracy-uncertainty relationship demonstrate the effectiveness of our method. Our code will be available at https://github.com/shizhediao/active-prompt.