S2LPP: Small-to-Large Prompt Prediction across LLMs

TL;DR

This paper investigates prompt preferences across various LLM sizes and proposes a cost-effective method for selecting prompts for larger models using smaller ones, demonstrating broad applicability and robustness.

Contribution

It introduces a novel approach leveraging prompt preference consistency across LLMs to reduce prompt engineering costs for large models.

Findings

Prompt preferences are consistent across different LLM sizes.

Using small models to select prompts for large models maintains performance.

The method applies effectively across multiple NLP tasks and fourteen LLMs.

Abstract

The performance of pre-trained Large Language Models (LLMs) is often sensitive to nuances in prompt templates, requiring careful prompt engineering, adding costs in terms of computing and human effort. In this study, we present experiments encompassing multiple LLMs variants of varying sizes aimed at probing their preference with different prompts. Through experiments on Question Answering, we show prompt preference consistency across LLMs of different sizes. We also show that this consistency extends to other tasks, such as Natural Language Inference. Utilizing this consistency, we propose a method to use a smaller model to select effective prompt templates for a larger model. We show that our method substantially reduces the cost of prompt engineering while consistently matching performance with optimal prompts among candidates. More importantly, our experiment shows the efficacy of our strategy across fourteen LLMs and its applicability to a broad range of NLP tasks, highlighting its robustness