A goal states what you want to accomplish. In a prompt, it tells the model which work matters instead of leaving it to guess from the topic alone. section 3.1.2
Consider “Tell me about electric cars.” The topic is clear, but the goal is not: you might want purchase costs, charging information, maintenance advice, or environmental impact. A useful goal begins with an action verb such as explain, compare, summarize, analyze, classify, generate, rewrite, translate, or plan. section 3.1 section 3.1.2
From topic to goal
“Tell me about electric cars” leaves the task open. “Compare electric cars with gasoline vehicles” gives the model a clear objective. Adding “in terms of purchase price, maintenance costs, driving range, and charging requirements” makes the desired work more precise. Context explains the situation; the goal says what to do; actions specify how the answer should look. section 3.1.2 section 3.1.3
Table 3.1 Common Main Points
|
Objective
|
Prompt template
|
|
Explain a concept
|
Explain [topic]
|
|
Compare alternatives
|
Compare [option A] and [option B]
|
|
Summarize information
|
Summarize [document/topic]
|
|
Analyze content
|
Analyze [data/text/document]
|
|
Classify information
|
Classify [item] according to [criteria]
|
|
Brainstorm ideas
|
Generate ideas for [goal]
|
|
Rewrite the content
|
Rewrite [text] according to [criteria]
|
|
Translate text
|
Translate [text] from [language A] to [language B]
|
|
Generate content
|
Create [content type] about [topic]
|
|
Plan a task
|
Create a plan for [goal]
|
Use the prompt templates to turn a broad intention into a concrete goal such as explaining, comparing, summarizing, analyzing, classifying, generating, rewriting, translating, or planning. section 3.1.2
A goal communicates intent rather than programming a fixed procedure. Traditional software specifies how a task must be performed, while prompting describes the outcome and lets the language model determine how to generate the response. section 3.1.3













