Google

An example template to directly instruct a Large Language Model to produce a prompt based on a task description is: `You are given a task to complete using LLMs. Please write a prompt to guide the LLMs. *task-description*`. This method provides a direct description of the task to the model.

Example of an Instruction for LLM-based Prompt Generation

Based on the scenario below, write a single, clear instruction that the marketing team could give to a general-purpose language model to help them create a high-quality prompt for their specific task.

Prompt Generation for a Specialized Task

A historian wants to use a large language model to analyze a collection of medieval manuscripts for mentions of previously unknown trade routes. The historian is an expert in history but has no experience in writing effective instructions for language models. Your task is to write a single, clear instruction that you would give to a powerful, general-purpose language model to help the historian. This instruction should ask the model to generate a detailed prompt that the historian can then use for their analysis.

Crafting a Prompt-Generating Instruction

A software development team wants to use a text-generating model to write user-friendly error messages for their application. However, the team lacks experience in crafting effective instructions for such a task. They decide to use a powerful, general-purpose model to help them create the necessary instructions first. Which of the following requests to the general-purpose model best demonstrates the strategy of using one model to generate instructions for another?

Learn Before

Related