The Intent Of The Prompt.

TL;DR:

In the context of Large Language Models (LLMs), understanding the intent of a prompt involves recognizing informative, persuasive, and transactional aspects in more complex prompts. Informative prompts seek to convey knowledge or facts, such as “Explain the process of photosynthesis.” Persuasive prompts aim to convince or influence, like “Argue for stricter environmental regulations.” Transactional prompts focus on completing a task or obtaining specific information, such as “Provide the nearest bus schedule.” By identifying the intent, users can craft prompts tailored to their desired outcome. For example, an informative prompt can encourage the LLM to provide comprehensive explanations, while a persuasive prompt prompts the LLM to present convincing arguments. Transactional prompts enable users to efficiently obtain the required information. Understanding the intent helps users achieve their communication goals by engaging the LLM in the appropriate manner and eliciting responses aligned with the desired prompt intent.

The intent of a prompt is the purpose or goal that the prompt is intended to achieve. It can vary depending on the context in which it is used, but generally it is meant to elicit a response or to guide the AI towards a particular action or decision. The intent of a prompt can be influenced by a variety of factors, including:

• the audience
• the desired outcome
• the tone of the prompt.

One important aspect to consider when crafting prompts is your audience (especially in a corporate context). Different audiences may have different expectations and preferences, and it is important to tailor the prompt to the needs and interests of the intended audience. Additionally, the desired outcome of the prompt should also be taken into account. Whether the goal is to persuade, inform, or entertain, the prompt should be crafted in a way that supports that goal.

The tone can influence how the AI perceives or rather interprets the intent of the prompt and so impacts the response. For example, a prompt that is written in a serious and informative tone may be more effective at conveying information, while a prompt that is written in a light-hearted and humorous tone may be more effective at eliciting an ’emotional response’. The AI will tell you Ad Infinitum it does not have emotions yet it will / may make an attempt to empathise with you.

Additional Factors

Other factors that can impact the intent of a prompt include cultural and societal norms, personal biases and beliefs, and the context in which the prompt is being used. For example, a prompt that is intended to elicit an emotional response may need to be crafted differently depending on the cultural and societal norms of the target audience. Additionally, personal biases and beliefs can impact how a reader interprets a prompt, it is important to be aware of these biases and to craft prompts in a way that is inclusive and respectful. In many cases you will find this ‘respectful language’ built-in to the AI.

Overall, the intent of a prompt is a critical aspect of prompt engineering and should be carefully considered in order to achieve the desired outcome. By taking into account the audience, desired outcome, tone and other factors that impact the intent of the prompt, prompt engineers can create prompts that are effective at eliciting a response and achieving their intended goal.

We offer specific ‘Intent’ related prompts in the final lesson of this course:

• Informative intent: “Can you explain how to change a tire on a car?”
• Persuasive intent: “Can you convince me that electric cars are a better choice than gas-powered cars?”
• Transactional intent: “Can you help me schedule an appointment with {my hairdresser} for next week?”