Based on the scenario below, what is the most likely reason for the model's undesirable behavior, and what specific instruction should be added to the beginning of the model's input to mitigate this issue?

Google

System information, also referred to as a system prefix, is a sequence of tokens positioned at the beginning of an input to guide the behavior of a Large Language Model. For example, it can be used to instruct the model to act as a helpful assistant and to refrain from producing toxic content.

System Information in LLM Inputs

An e-commerce company wants to use a large language model as a customer service chatbot. The model must be polite, stay on the topic of the company's products, and decline to answer unrelated questions or invent information. Which of the following initial instructions would be most effective at guiding the model to meet all these requirements?

Troubleshooting an LLM-Powered Internal Knowledge Base

You are tasked with configuring a large language model to act as a tutor for 5th-grade math students. The model's responses must be encouraging, break down problems into simple steps, and avoid giving the final answer directly, instead guiding the student to discover it themselves. Write a 'system information' prompt that would guide the model to adopt this specific persona and set of behaviors.

Learn Before

Related