Large Language Models as Sous Chefs: Revising Recipes with GPT-3

TL;DR

This paper demonstrates how large language models, especially GPT-3.5, can effectively revise and simplify recipes from various cuisines, serving as digital sous chefs to improve usability and understanding.

Contribution

The study introduces a novel prompt design for recipe revision using LLMs, evaluates multiple models, and provides a human judgment framework, releasing tools for public use.

Findings

Annotators prefer revised recipes over original ones

GPT-3.5 achieves the best revision quality

The approach is applicable across diverse cuisines

Abstract

With their remarkably improved text generation and prompting capabilities, large language models can adapt existing written information into forms that are easier to use and understand. In our work, we focus on recipes as an example of complex, diverse, and widely used instructions. We develop a prompt grounded in the original recipe and ingredients list that breaks recipes down into simpler steps. We apply this prompt to recipes from various world cuisines, and experiment with several large language models (LLMs), finding best results with GPT-3.5. We also contribute an Amazon Mechanical Turk task that is carefully designed to reduce fatigue while collecting human judgment of the quality of recipe revisions. We find that annotators usually prefer the revision over the original, demonstrating a promising application of LLMs in serving as digital sous chefs for recipes and beyond. We release our prompt, code, and MTurk template for public use.