A developer's goal is to use an AI system to understand the connections between key pieces of information in a text. Analyze the following scenario and explain why the provided instruction was ineffective for achieving this goal. Then, propose a more effective instruction.

Google

A prompt for a relation extraction task instructs a system to identify and describe the contextual connections between pre-identified named entities in a text. A clear example of such a prompt is: 'Given a text and a list of named entities identified within it, analyze and describe the relationships between these entities. Explain how each entity is contextually related to others.' This task requires the system to go beyond simple entity listing and interpret the semantic links that connect them.

Example of a Relation Extraction Task Prompt

A user wants to determine the specific connection between 'Jane Doe' and 'Innovate Corp' in the following sentence: 'Innovate Corp, headquartered in Silicon Valley, announced that its founder, Jane Doe, will be stepping down as CEO.' Which of the following instructions would be most effective for this specific task?

Evaluating Instructional Effectiveness for Text Analysis

You are given the sentence: "Dr. Anya Sharma, a lead researcher at the Global Health Institute, published a groundbreaking study on malaria vaccines in the 'Journal of Medicine'." The named entities identified are 'Dr. Anya Sharma', 'Global Health Institute', and 'Journal of Medicine'. Write a clear and concise instruction for a language model to identify and describe the relationships between these three entities based *only* on the provided sentence.

Learn Before

Related