論文研究關鍵字

首頁Home › 人工智慧與新興科技 AI & Emerging Tech

NO.806

提示詞工程

Prompt Engineering

1這是什麼What is it?

提示詞工程(prompt engineering)是設計與調整輸入指令以引導生成模型產出預期結果的實踐。在圖像生成中涉及主體、風格、構圖、光線、媒材與負面提示的組織;在語言模型中則有零樣本、少樣本與思維鏈(chain-of-thought)等技巧。其方法多為經驗性,且隨模型版本而變動,缺乏穩定的理論基礎。

Prompt engineering is the practice of designing and adjusting input instructions to steer generative models toward intended results. In image generation it involves organizing subject, style, composition, lighting, medium and negative prompts, and in language models there are techniques such as zero-shot, few-shot and chain-of-thought, its methods are largely empirical and vary with model versions, lacking a stable theoretical basis.

2常被用在哪種研究問題Typical research questions

探索型Exploratory影響/關聯型Effect / association

常見問法:提示詞策略如何影響生成結果?

示範寫法Example

提示詞撰寫是新的設計技能。本研究觀察設計者撰寫與修改提示詞的歷程,並歸納有效的策略。

Prompt writing is a new design skill. This study observes how designers write and revise prompts and summarizes effective strategies.

撰寫內容,非真實研究結果,請與指導教授討論後改寫。

Written as a model example, not a real finding. Please revise with your advisor.

3怎麼測它How to measure it

提示詞操弄+結果評分

Measurement notes are written in Chinese. Switch to 中/EN to read them.

4代表文獻Key references

英文論文(2000 年後)English paper (post-2000)

Liu, V., & Chilton, L. B. (2022). Design guidelines for prompt engineering text-to-image generative models. CHI '22, ACM.

中文論文(2000 年後)

中文論文為預填檢索連結,請自行挑選一篇 2000 年後的期刊論文並核對作者、年份、出處後再引用。

英文引用來自模型整理,引用前請務必核對並閱讀原文。English citations are compiled by a model. Verify and read the original before citing.

‹ 文字生成圖像Text-to-ImageAI輔助設計 ›AI-assisted Design