Google

A complete setup for a relationship extraction task includes an instructional prompt, a source text, and a list of named entities. For example, a system could be given the prompt: 'Given a text and a list of named entities identified within it, analyze and describe the relationships between these entities.' This instruction would be paired with a source text, such as the 2024 tourism outlook, and a specific list of entities to analyze from that text: 'Tom Jenkins, European Tourism Organisation, UK, Europe, 2024'.

Example of a Relationship Extraction Task Prompt

A researcher wants to use a language model to understand the connections between specific people and organizations mentioned in a news snippet. Which of the following prompts is best structured to perform this task?

You are a data scientist tasked with extracting structured information from a news article. Your goal is to identify the relationships between key entities mentioned in a specific sentence. Based on the provided text and list of entities below, construct a complete and effective prompt for a large language model to perform this relationship extraction task.

Crafting an Effective Relationship Extraction Prompt

You are setting up a task for a language model to identify how different elements in a document are connected. Match each component of the task setup with its correct description.

Learn Before

Related