Instances Need More Care: Rewriting Prompts for Instances with LLMs in the Loop Yields Better Zero-Shot Performance

TL;DR

This paper introduces PRomPTed, a novel method that rewrites prompts for individual instances using LLMs in the loop, significantly improving zero-shot performance across multiple datasets and models, including open-source LLMs.

Contribution

The paper proposes PRomPTed, an innovative prompt rewriting approach that enhances zero-shot task performance by optimizing prompts at the instance level using LLMs in the loop.

Findings

PRomPTed outperforms naive zero-shot and output refinement baselines.

Prompt rewriting with GPT-3.5 can match or exceed GPT-4's performance.

Advantages of prompt rewriting generalize to open-source LLMs.

Abstract

Large language models (LLMs) have revolutionized zero-shot task performance, mitigating the need for task-specific annotations while enhancing task generalizability. Despite its advancements, current methods using trigger phrases such as "Let's think step by step" remain limited. This study introduces PRomPTed, an approach that optimizes the zero-shot prompts for individual task instances following an innovative manner of "LLMs in the loop". Our comprehensive evaluation across 13 datasets and 10 task types based on GPT-4 reveals that PRomPTed significantly outperforms both the naive zero-shot approaches and a strong baseline (i.e., "Output Refinement") which refines the task output instead of the input prompt. Our experimental results also confirmed the generalization of this advantage to the relatively weaker GPT-3.5. Even more intriguingly, we found that leveraging GPT-3.5 to rewrite prompts for the stronger GPT-4 not only matches but occasionally exceeds the efficacy of using GPT-4 as the prompt rewriter. Our research thus presents a huge value in not only enhancing zero-shot LLM performance but also potentially enabling supervising LLMs with their weaker counterparts, a capability attracting much interest recently. Finally, our additional experiments confirm the generalization of the advantages to open-source LLMs such as Mistral 7B and Mixtral 8x7B.