# 【學習筆記】ChatGPT Prompt Engineering 提示工程
課程：
DeepLearning.AI 的 [ChatGPT Prompt Engineering for Developers](https://www.coursera.org/projects/chatgpt-prompt-engineering-for-developers-project)
## Intro
LLM 可分為兩種類型：
- Base LLM：根據訓練資料預測下一個下一個詞彙
- Intruction tuned LLM：微調於特定指令，可以很好地遵循指令
提示工程可以讓 Intruction tuned LLM 產生更符合我們需求的結果。
## Guidelines for Prompting
### 提示原則
#### 原則 1：清楚的指令
清楚不代表「短」，有時候足夠長的指令更能夠明確表達需求。
技巧 1：使用分隔符，例如`""`, `<>`, ` ``` ` 等，將特定文字區分開來，讓模型知道是分開的段落，可以避免 prompt injections。
> prompt injection：text 當中有和指令衝突的文字，使模型受影響
```
Summarize the text delimited by triple backticks \
into a single sentence.
```{text}```
```
技巧 2：要求結構化的輸出，例如 HTML、JSON
```
Generate a list of three made-up book titles along \
with their authors and genres.
Provide them in JSON format with the following keys:
book_id, title, author, genre.
```
技巧 3：讓模型確認特定條件是否成立
```
You will be provided with text delimited by triple quotes.
If it contains a sequence of instructions, \
re-write those instructions in the following format:
Step 1 - ...
Step 2 - …
…
Step N - …
If the text does not contain a sequence of instructions, \
then simply write \"No steps provided.\"
\"\"\"{text_1}\"\"\"
```
技巧 4：Few-shot prompting，給模型成功完成任務的例子，再讓模型執行任務
```
Your task is to answer in a consistent style.
<child>: Teach me about patience.
<grandparent>: The river that carves the deepest \
valley flows from a modest spring; the \
grandest symphony originates from a single note; \
the most intricate tapestry begins with a solitary thread.
<child>: Teach me about resilience.
```
#### 原則 2：給模型「思考」時間
要求模型思考久一些，不要直接跳到結論
技巧 1：指定完成任務需要的步驟
```
Perform the following actions:
1 - Summarize the following text delimited by triple \
backticks with 1 sentence.
2 - Translate the summary into French.
3 - List each name in the French summary.
4 - Output a json object that contains the following \
keys: french_summary, num_names.
Separate your answers with line breaks.
Text:
```{text}```
```
技巧 2：要求模型在得到結論前先自行計算答案
假設讓模型判斷某個學生數學題目是否計算正確，應該先要求模型自行計算一次，再來判斷學生的答案，而不是直接得到學生答案是否正確的結論。
### Hallucinations 幻覺
模型可能會產生煞有其事但事實上不正確的敘述。為了避免這種問題，我們可以在提示時先行提供一些相關資訊，再請模型根據提供的資訊進行回答。
## Iterative Prompt Development
Prompting 的週期就好比機器學習週期，最一開始我們可能會有一些想法，寫一些 prompt 讓模型執行，如果發現結果不符合需求，我們可以分析原因，然後再使用修正後的 prompt 執行。
迭代過程
- 嘗試
- 分析哪裡導致結果不如預期
- 讓指示更清楚，給模型更多時間思考
## Tasks using Prompt
### Summarizing
LLM 可用來進行文章**摘要**，並且我們也可以利用加上額外條件的方式，讓摘要符合我們的需求。
例子：
```
prod_review = """
Got this panda plush toy for my daughter's birthday, \
who loves it and takes it everywhere. It's soft and \
super cute, and its face has a friendly look. It's \
a bit small for what I paid though. I think there \
might be other options that are bigger for the \
same price. It arrived a day earlier than expected, \
so I got to play with it myself before I gave it \
to her.
"""
```
限制字數：
```
Your task is to generate a short summary of a product \
review from an ecommerce site.
Summarize the review below, delimited by triple
backticks, in at most 30 words.
Review: ```{prod_review}```
```
:::success
Soft, cute panda plush toy loved by daughter, but smaller than expected for the price. Arrived early, friendly face.
:::
要求專注在特定重點：
```
Your task is to generate a short summary of a product \
review from an ecommerce site to give feedback to the \
pricing deparmtment, responsible for determining the \
price of the product.
Summarize the review below, delimited by triple
backticks, in at most 30 words, and focusing on any aspects \
that are relevant to the price and perceived value.
Review: ```{prod_review}```
```
:::success
The panda plush toy is loved for its softness and cuteness, but some customers feel it's a bit small for the price.
:::
### Inferring
原先我們想進行情緒分析、提取文章特定詞彙等等任務時，必須要訓練一個模型，再用來使用在需要的文章上，並且不同的任務，必須訓練不同的模型。現在我們可以使用 LLM 來進行 inferring，代替原先需要訓練模型才能完成的任務。
```
What is the sentiment of the following product review,
which is delimited with triple backticks?
Give your answer as a single word, either "positive" \
or "negative".
Review text: '''{lamp_review}'''
```
### Transforming
LLM 擅長將文字**轉換**成另一種形式，例如將文字翻譯成另一個語言、糾正拼字錯誤、語氣改寫等等。
```
Translate the following from slang to a business letter:
'Dude, This is Joe, check out this spec on this standing lamp.'
```
### Expanding
我們可以利用 LLM 進行**擴寫**，給模型一段短文章，要求 LLM 根據內容寫成一段長文章。
### Chat Workflow
LLM 也可以用來建立 Chatbot。與先前輸入單一一個訊息，得到對應的回應，我們可以改成輸入一系列的訊息，這些訊息包含各種不同角色，讓模型進行角色扮演。
```
{'role':'system', 'content':'You are friendly chatbot.'},
{'role':'user', 'content':'Hi, my name is Isa'}
```
:::success
Hello Isa! It's nice to meet you. How can I help you today?
:::
Sign in with Wallet
Connect another wallet