PAL（程序辅助语言模型）
PAL（程序辅助语言模型）
PAL（程序辅助语言模型） PAL（程序辅助语言模型） Modified November 26, 2025 Gao 等人（2022）提出了一种使用 LLMs 读取自然语言问题并生成程序作为中间推理步骤的方法。被称为程序辅助语言模型（PAL），它与思维链提示不同，因为它不是使用自由形式文本来获得解决方案，而是将解决步骤卸载到类似 Python 解释器的编程运行时中。 图片来源：Gao 等人（2022） 让我们以 LangChain 和 OpenAI GPT 3 为例。我们有兴趣开发一个简单的应用程序，它能够解释所提出的问题，并利用 Python 解释器提供答案。 具体来说，我们有兴趣创建一个功能，允许使用 LLM 回答需要日期理解的问题。我们将为 LLM 提供一个提示，其中包括一些示例，这些示例是从这里采用的。 这是我们需要导入的包： Code block Plain Text import openaifrom datetime import datetimefrom dateutil.relativedelta import relativedeltaimport osfrom langchain.llms import OpenAIfrom dotenv import load dotenv 让我们先配置一些环境： Code block Plain Text load dotenv() API configurationopenai.api key = os.getenv("OPENAI API KEY") for LangChainos.environ["OPENAI API KEY"] = os.getenv("OPENAI API KEY") 设置模型实例： Code block Plain Text llm = OpenAI(model name='text davinci 003', temperature=0) 设置提示+问题： Code block Shell question = "Today is 27 February 2023. I was born exactly 25 years ago. What is the date I was born in MM/DD/YYYY?" DATE UNDERSTANDING PROMPT = """ Q: 2015 is coming in 36 hours. What is the date one week from today in MM/DD/YYYY? If 2015 is coming in 36 hours, then today is 36 hours before. today = datetime(2015, 1, 1) relativedelta(hours=36) One week from today, one week from today = today + relativedelta(weeks=1) The answer formatted with %m/%d/%Y is one week from today.strftime('%m/%d/%Y') Q: The first day of 2019 is a Tuesday, and today is the first Monday of 2019. What is the date today in MM/DD/YYYY? If the first day of 2019 is a Tuesday, and today is the first Monday of 2019, then today is 6 days later. today = datetime(2019, 1, 1) + relativedelta(days=6) The answer formatted with %m/%d/%Y is today.strftime('%m/%d/%Y') Q: The concert was scheduled to be on 06/01/1943, but was delayed by one day to today. What is the date 10 days ago in MM/DD/YYYY? If the concert was scheduled to be on 06/01/1943, but was delayed by one day to today, then today is one day later. today = datetime(1943, 6, 1) + relativedelta(days=1) 10 days ago, ten days ago = today relativedelta(days=10) The answer formatted with %m/%d/%Y is ten days ago.strftime('%m/%d/%Y') Q: It is 4/19/1969 today. What is the date 24 hours later in MM/DD/YYYY? It is 4/19/1969 today. today = datetime(1969, 4, 19) 24 hours later, later = today + relativedelta(hours=24) The answer formatted with %m/%d/%Y is today.strftime('%m/%d/%Y') Q: Jane thought today is 3/11/2002, but today is in fact Mar 12, which is 1 day later. What is the date 24 hours later in MM/DD/YYYY? If Jane thought today is 3/11/2002, but today is in fact Mar 12, then today is 3/12/2002. today = datetime(2002, 3, 12) 24 hours later, later = today + relativedelta(hours=24) The answer formatted with %m/%d/%Y is later.strftime('%m/%d/%Y') Q: Jane was born on the last day of Feburary in 2001. Today is her 16 year old birthday. What is the date yesterday in MM/DD/YYYY? If Jane was born on the last day of Feburary in 2001 and today is her 16 year old birthday, then today is 16 years later. today = datetime(2001, 2, 28) + relativedelta(years=16) Yesterday, yesterday = today relativedelta(days=1) The answer formatted with %m/%d/%Y is yesterday.strftime('%m/%d/%Y') Q: {question} """.strip() + '\n' Code block Plain Text llm out = llm(DATE UNDERSTANDING PROMPT.format(question=question))print(llm out) 这将输出以下内容： Code block Shell If today is 27 February 2023 and I was born exactly 25 years ago, then I was born 25 years before. today = datetime(2023, 2, 27) I was born 25 years before, born = today relativedelta(years=25) The answer formatted with %m/%d/%Y is born.strftime('%m/%d/%Y') llm out 是一段 python 代码，我们可以使用 exec 执行它： Code block Plain Text exec(llm out)print(born) 这将输出以下内容：02/27/1998 You don't have permission to access this synced block. Please log in and try again. Log In Gao 等人（2022） Gao 等人（2022） 这里 Gao 等人（2022）提出了一种使用 LLMs 读取自然语言问题并生成程序作为中间推理步骤的方法。被称为程序辅助语言模型（PAL），它与思维链提示不同，因为它不是使用自由形式文本来获得解决方案，而是将解决步骤卸载到类似 Python 解释器的编程运行时中。 Gao 等人（2022） 图片来源：Gao 等人（2022） Gao 等人（2022） 让我们以 LangChain 和 OpenAI GPT 3 为例。我们有兴趣开发一个简单的应用程序，它能够解释所提出的问题，并利用 Python 解释器提供答案。 具体来说，我们有兴趣创建一个功能，允许使用 LLM 回答需要日期理解的问题。我们将为 LLM 提供一个提示，其中包括一些示例，这些示例是从这里采用的。 这里 这是我们需要导入的包： 让我们先配置一些环境： 设置模型实例： 设置提示+问题： 这将输出以下内容： llm out 是一段 python 代码，我们可以使用 exec 执行它： 这将输出以下内容：02/27/1998 You don't have permission to access this synced block. Please log in and try again. Log In