場景 7：化繁為簡——訊息總結
場景介紹
內容生成大場景下的第二個場景是化繁為簡，這個場景其實很好理解，就是將複雜的內容，轉為簡單的內容，一般常遇到的場景有：
- 訊息總結：顧名思義，就是對一堆訊息進行總結。
- 訊息解釋：這個跟改寫內容有點像，但這個更偏向於解釋與總結。下一章會給大家介紹更多的例子。
- 訊息提取：提取訊息裡的某一段內容，比如從一大段文字中，找到關鍵內容，並分類。
本章會講一下訊息總結。訊息總結還是比較簡單的，基本上在 prompt 里加入總結 summarize 就可以了。但如果你想要一些特別的效果，不妨組合使用之前介紹的技巧，比如：
- 使用技巧 2，增加總結範例，讓 AI 總結符合你需求的內容
- 使用技巧 4，增加 role，讓 AI 總結的內容具有一定的風格
不過在這個場景，還有個技巧需要各位注意。
技巧 5：使用特殊符號將指令和需要處理的文字分開
不管是訊息總結，還是訊息提取，你一定會輸入大段文字，甚至多段文字，此時有個小技巧。
可以用『』『將指令和文字分開。根據我的測試，如果你的文字有多段，增加』『』會提升 AI 反饋的準確性（這個技巧來自於 OpenAI 的 API 最佳實踐文件）
info
感謝 CraneHuang6 的提醒，這裡還能用 ### 符號區隔，不過我一般會用『』『 ，因為我有的時候會用 # 作為格式範例，太多 # 的話 prompt 會看起來比較暈 😂
像我們之前寫的 prompt 就屬於 Less effective prompt。為什麼呢？據我的測試，主要還是 AI 不知道什麼是指令，什麼是待處理的內容，用符號分隔開來會更利於 AI 區分。
Please summarize the following sentences to make them easier to understand.
OpenAI is an American artificial intelligence (AI) research laboratory consisting of the non-profit OpenAI Incorporated (OpenAI Inc.) and its for-profit subsidiary corporation OpenAI Limited Partnership (OpenAI LP). OpenAI conducts AI research with the declared intention of promoting and developing a friendly AI. OpenAI systems run on the fifth most powerful supercomputer in the world.[5][6][7] The organization was founded in San Francisco in 2015 by Sam Altman, Reid Hoffman, Jessica Livingston, Elon Musk, Ilya Sutskever, Peter Thiel and others,[8][1][9] who collectively pledged US$1 billion. Musk resigned from the board in 2018 but remained a donor. Microsoft provided OpenAI LP with a $1 billion investment in 2019 and a second multi-year investment in January 2023, reported to be $10 billion.[10]
Better prompt:
Please summarize the following sentences to make them easier to understand.
Text: """
OpenAI is an American artificial intelligence (AI) research laboratory consisting of the non-profit OpenAI Incorporated (OpenAI Inc.) and its for-profit subsidiary corporation OpenAI Limited Partnership (OpenAI LP). OpenAI conducts AI research with the declared intention of promoting and developing a friendly AI. OpenAI systems run on the fifth most powerful supercomputer in the world.[5][6][7] The organization was founded in San Francisco in 2015 by Sam Altman, Reid Hoffman, Jessica Livingston, Elon Musk, Ilya Sutskever, Peter Thiel and others,[8][1][9] who collectively pledged US$1 billion. Musk resigned from the board in 2018 but remained a donor. Microsoft provided OpenAI LP with a $1 billion investment in 2019 and a second multi-year investment in January 2023, reported to be $10 billion.[10]
"""
另外，在吳恩達的 ChatGPT Prompt Engineering 課程中，還提到，你可以使用其他特殊符號來分割文字和 prompt，比如<>，<tag></tag> 等，課程中的案例是這樣的（注意這個是 python 程式碼，需要關注的是 prompt 裡的 text）：
text = f"""
You should express what you want a model to do by \
providing instructions that are as clear and \
specific as you can possibly make them. \
This will guide the model towards the desired output, \
and reduce the chances of receiving irrelevant \
or incorrect responses. Don't confuse writing a \
clear prompt with writing a short prompt. \
In many cases, longer prompts provide more clarity \
and context for the model, which can lead to \
more detailed and relevant outputs.
"""
prompt = f"""
Summarize the text delimited by triple backticks \
into a single sentence.
`{text}`
"""
如果你是在開發一款允許使用者輸入內容，並對內容進行總結的 AI 產品，那你就可以使用到這個技巧。