当前位置：首页 > news >正文

(undone) 吴恩达版提示词工程 2. 指南

news 来源：原创 2025/4/21 14:21:14

url: https://www.bilibili.com/video/BV1Z14y1Z7LJ?spm_id_from=333.788.videopod.episodes&vd_source=7a1a0bc74158c6993c7355c5490fc600&p=2

别人的笔记 url: https://zhuanlan.zhihu.com/p/626966526

指导原则（Guidelines）

编写提示词有两个原则：
1.编写具体、明确的提示词
2.给模型足够的时间思考

1. 系统配置

如下是使用 openai API 的方法：
安装模块

python3 -m pip install openai
python3 -m pip install dotenv

在 openai 官网创建 API key，充值，随后在自己的 Linux 电脑上设置环境变量：

export OPENAI_API_KEY=<openai-api-key>

使用下面的代码能迅速检验自己是否能使用 openai api key：

from openai import OpenAIclient = OpenAI()response = client.responses.create(model="gpt-4.1",input="Write a one-sentence bedtime story about a unicorn."
)print(response.output_text)

我们将使用 OpenAI 的 GPT 3.5 Turbo模型，并使用 chat completion API。我们将在稍后的视频中详细介绍 chat completion API 的格式和输入。

现在我们只要定义一个辅助函数 get_completion() ，以便使用提示和查看生成的输出。函数 get_completion() 接收一个提示 prompt，返回该提示的完成内容。

 def get_completion(prompt, model="gpt-3.5-turbo"):messages = [{"role": "user", "content": prompt}]response = openai.ChatCompletion.create(model=model,messages=messages,temperature=0, # this is the degree of randomness of the model's output)return response.choices[0].message["content"]

2. 指导原则 1：清晰而具体的提示

现在，让我们讨论提示的第一个指导原则，是编写清晰而具体的提示。

你应该提供尽可能清晰而具体的说明，来表达你希望模型执行的任务。这将指导模型生成期望的输出，减少无关或错误响应的可能。

不要把清晰的提示和简短的提示混为一谈。在很多情况下，较长的提示可以为模型提供更多的清晰度和上下文，从而产生更详细和更相关的输出。

第一个策略：使用分隔符来清楚地表示输入的不同部分

我来举个例子。我们有一段话，我们想要完成的任务就是总结这段话。因此，我在提示中要求，将由三重反引号```分隔的文本总结为一句话。

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}```
"""
response = get_completion(prompt)
print(response)

在提示中，我们使用三重反引号```把将文本{text}括起来，使用 get_completion 函数获得响应，然后打印输出响应。如果我们运行这段程序，就可以得到下面这个输出的句子。

下面是总代码：
注意，下面的代码很多 API 过时，已经被废弃，需要使用新的代码

import openai
import os# 1. 根据环境变量获取 openai key
from dotenv import load_dotenv, find_dotenv
_ = load_dotenv(find_dotenv())openai.api_key = os.getenv('OPENAI_API_KEY') # 2. 定义 get_completion 方法
def get_completion(prompt, model="gpt-3.5-turbo"):messages = [{"role": "user", "content": prompt}]response = openai.ChatCompletion.create(model=model,messages=messages,temperature=0, # this is the degree of randomness of the model's output)return response.choices[0].message["content"] # 3. 使用模型
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}```
"""response = get_completion(prompt)print(response)

新的代码如下：

import os
from openai import OpenAI# 1. 根据环境变量获取 openai key
client = OpenAI(# This is the default and can be omittedapi_key=os.environ.get("OPENAI_API_KEY"),
)# 2. 定义 get_completion 方法
def get_completion(prompt, model="gpt-3.5-turbo"):response = client.responses.create(model=model,instructions=f"""Summarize the text delimited by triple backticks \ into a single sentence.""",input=prompt,temperature=0, # this is the degree of randomness of the model's output)return response.output_text# 3. 使用模型
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"""
```{text}```
"""response = get_completion(prompt)print(response)

It is important to provide clear and specific instructions
to guide a model towards the desired output, as longer
prompts can offer more clarity and context for the model, 
resulting in more detailed and relevant responses.

在本例中我们使用这些分隔符，向模型非常清楚地指定它应该使用的确切文本。

分隔符可以是任何明确的标点符号，将特定的文本片段部分与提示的其它部分分隔开来。分隔符可以使用三重双引号、单引号、XML标记、章节标题，或者任何可以向模型表明这是一个单独部分的符号或标记。例如我们可以使用这些分隔符： “”"，—，< >，。