English 繁體中文 简体中文

AI 代理如何处理追问(ReAct,版本 5,无记忆)?

林邦源博士 · AI 笔记 · 2026 年 11 月 9 日 · 约 3 分钟阅读

版本 4把代理变成了一 个可以重复使用的函数。版本 5 连续调用它两次:先问“Mark 何时推出 Muse?”, 然后是一个自然的追问,“香港现在能用了吗?”。人类读者一听就知“它”指 什么,代理却不知道 —— 而看着它发现这一点,正是这一课。

脚本

这是 ReAct05.py。除了最后两行,每一行都与版本 4 完全相同 —— 现在有两次调用:

def perform_task(task):
    """Run the Think -> Act -> Observe loop for one task and return the answer."""
    prompt = f"""Task: {task}
You can use ONE tool: search[query] searches the web (Tavily) and returns the top results.
Today is {datetime.now().strftime('%Y-%m-%d')}.
Work in steps. Each step has a Thought, an Action, and an Observation."""

    for step in range(1, 6):          # at most 5 rounds, so it cannot loop forever
        thought = ask(prompt + f"\n\nThought {step}: (reason briefly about what to do next; do not act yet)",
                      stop=["Action"])
        print(f"=== Round {step} - Thought ===")
        print(thought)

        prompt = prompt + f"\n\nThought {step}: {thought}"
        action = ask(prompt + "\n\nWhat action should be taken based on that? "
                     "Reply with ONLY: search[query] or finish[answer]").strip()
        print(f"=== Round {step} - Action chosen by the model ===")
        print(action)

        if action.startswith("finish"):
            found = re.search(r"finish\[(.*)\]", action, re.S)
            answer = found.group(1) if found else action
            print("=== Final Answer ===")
            print(answer)
            return answer

        query = re.search(r"search\[(.*)\]", action, re.S)
        observation = search_tavily(query.group(1)) if query else "Invalid action."
        print(f"=== Round {step} - Observation from Tavily web search ===")
        print(observation)
        prompt = prompt + f"\nAction {step}: {action}\nObservation {step}: {observation}"
    else:
        print("=== Stopped: no answer after 5 rounds ===")

perform_task("When did Mark launch Muse")
perform_task("Is it available in Hong Kong yet?")

文件顶部的 import、两个 API 客户端、ask() 和 search_tavily() 都与 版本 3相同,所以这里 省略。

代码中唯一要看的地方

看 perform_task 内的第一行:每次调用这个函数,提示词都是 从零开始建立的。第一次调用的任何东西 —— 问题、搜索、答案 —— 都不会留到第二次。函数可以重复使用,但同时也是无状态的: 两次调用就是两个陌生人。

一次真实运行

--- Task 1: "When did Mark launch Muse" ---
...four rounds of searching; sources disagree by one day...
=== Final Answer ===
Meta (Mark Zuckerberg) launched Muse, its personal AI agent, on September 8, 2026 (with some outlets citing September 9, 2026).

--- Task 2: "Is it available in Hong Kong yet?" ---
=== Round 1 - Thought ===
Thought 1: The query is ambiguous because "it" isn't specified. I need to figure out what product or service the user is asking about. Since the user only wrote "Is it available in Hong Kong yet?" with no prior context, I don't know the referent... The correct move in a real scenario is to ask the user. But the format forces me to do a search. I'll do a broad search on "available in Hong Kong 2026"...
=== Round 1 - Action chosen by the model ===
search[available in Hong Kong 2026]
=== Round 1 - Observation from Tavily web search ===
[Hong Kong 2026 | CanJam] ...headphone festival, tickets...
[Hong Kong Eats 2026 - Hungry Onion] ...restaurant chat...
[Hong Kong Conditions Report 2026 - US Department of State] ...policy report...
=== Round 2 - Thought ===
Thought 2: The search results ... don't reveal what product or service the user is asking about... I cannot determine the answer. I should state that clarification is needed...
=== Round 2 - Action chosen by the model ===
finish[I can't determine whether "it" is available in Hong Kong because the item/service isn't specified. Please tell me what "it" refers to (e.g., a product, app, game, streaming service, payment feature), and I'll check availability in Hong Kong.]

节选已缩短(…);完整运行会打印出两个任务的全部内容。

任务 1 一如以往运行正常。任务 2 才是有趣的失败:模型立刻察觉到含糊之处,用一次 搜索证明盲目查询只会换来耳机节和签证报告,然后用 finish[] 做唯一 诚实的事 —— 要求澄清。这个行为值得欣赏,但设计上的缺陷在我们身上:我们向一个 记不起第一个问题的代理,问了一个追问。

补救方法:把上下文交给代理

有两种做法。

做法一:自己带着上下文

在这个层次,“记忆”只是你一路传递的文字。你可以把两个问题合并成一个任务:

perform_task("When did Mark launch Muse, and is it available in Hong Kong yet?")

或者把第一个答案喂回第二个任务:

answer1 = perform_task("When did Mark launch Muse")
perform_task(f"Context: {answer1} Is it available in Hong Kong yet?")

第二种写法就是聊天机器人运行的雏形:它们好像记得整段对话,其实只是应用程序 每收到一条消息,都悄悄地把整段历史重新发送出去。

做法二:给代理记忆

手动传递文字,应付两个问题还可以,但应付不了一场真实对话。正确的补救方法是让 代理拥有自己的记忆,这样每次调用都能看到之前发生过什么。这就是这个系列下一篇 AI 笔记的主题。

在自己的电脑上运行

下载演示程序 (Folder-AI-ReAct05.zip)。zip 内有三个文件: ReAct05.py(完整脚本)、pyproject.toml 和 README.md。

  1. 在 Visual Studio Code 打开这个文件夹。
  2. 在文件夹内创建一个名为 .env 的文件,写上两行:
    DEEPSEEK_API_KEY=your-deepseek-key
    TAVILY_API_KEY=your-tavily-key
    密钥:DeepSeek 控制台和 Tavily (两者都有免费的起始额度)。
  3. 在终端运行:
    uv sync
    uv run ReAct05.py
    然后试试上面两种补救方法,比较任务 2 的回答有何不同。

为什么这个实验值得关注

延伸阅读