5 個常見提示詞錯誤及修正方法
5 Common Prompt Mistakes (and How to Fix Them)
提示詞常見的 5 個錯誤包括描述含糊、缺少背景、未指定格式、一次提出太多問題,以及沒有交代回答對象。修正時應明確説明任務與情境、指定輸出格式、一次集中處理一個問題,並説明答案是寫給誰看。
Why your prompt matters
Prompt is what you gave to AI chatbot or the question or instruction you gave to AI to get a response out of it.
Vague or clumsy prompt will give vague or clumsy response. The clearer prompt the better the response.
Let's look at some common mistakes.
Mistake 1: Being too vague
Before: Explain a maths theorem.
After: Explain Pythagoras theorem to a ten year old kid, with simple everyday example.
Why this works: The AI knows which topic to cover and how to explain it.
Mistake 2: Giving no context
Before: Write a cover letter
After: Write a cover letter for an engineer with 4+ years experience in web development, transitioning into tech writing.
Why this works: The AI knows who you are, what you do and can reply in a way that fits your situation.
Mistake 3: Not specifying the format
Before: Give me yoga postures.
After: Give me 5 yoga postures for begineers in bulleted list with one sentence each.
Why this works: AI provides response in a specific format which user asked, response is easy and short to understand.
Mistake 4: Asking too many things at once
Before: Why is the sun hot? Why is earth round? How many galaxies are in the entire universe? How many black holes are discovered? What is a black hole?
After: Why is sun hot?
Why this works: AI chatbot can focus on one question, instead of getting confused or giving rushed response for each.
Mistake 5: Not telling the AI who it’s talking to
Before: What's perimeter of a rectangle?
After: Assume you are talking to a 12-year-old. Explain the perimeter of a rectangle to a 12-year-old using a simple example.
Why this works: AI chatbot focuses on simple answer, choosing examples which fit the reader better.
Conclusion
Better prompt means better responses from chatbot. Be specific, give context, name the format, ask one thing at a time and say who the answer is for.
來源:DEV Community · dev.to