

Who this is for
- Developers who need to deal with quotations, statistics, and formula calculations
- Readers who are sensitive to model hallucinations
What problem does this article solve?
Let the model be error-prone and lack a verifiable calculation path.
Key concepts
- tool division
- parametric extraction
- Validated calculations
- result explanation
- Error input processing
Walkthrough
- Design Calculator tool
- parsing expression parameters
- Use safe calculation functions
- Let the model interpret the calculation results
Python code example
The companion demo is generated in:
cd examples/llm-guides/014_tool_calling_calculator
python demo.py
The example follows two principles:
- From
OPENAI_API_KEY, load the key at runtime to avoid hard-coding it. - Without an API key, the example runs in mock mode so you can follow the workflow.
The essential call looks like this:
import os
api_key = os.getenv("OPENAI_API_KEY")
if not api_key:
print("未配置 OPENAI_API_KEY,进入 mock 模式。")
else:
# 真实项目中在这里初始化 SDK 客户端并发起请求。
print("已读取 OPENAI_API_KEY,可以调用模型。")
Common pitfalls
- Publishing after a single successful demo run leaves out error handling, logging and human review.
- Including API keys, real file paths, account details or business data in published articles or code.
- Readers can struggle to reproduce an example if it is unclear which parts are mocked and which run in production.
- Citing an API or framework feature without checking its documentation can mislead readers when versions change.
Summary
- LLM Guide 14 shows how to use a calculator to avoid confidently wrong answers to math questions. When a model attempts mental arithmetic, mistakes are easy to make and the calculation is hard to verify.
- Make the workflow reproducible, reversible and easy to review. One good generation tells you little about how it will hold up in use.
- Before publishing, check current documentation for model capabilities, API parameters and third-party frameworks.
CSDN tags
Large model, Python, Tool Calling, Calculator, Reliability
WeChat article summary
The calculator illustrates a useful division of labor. When a task has a verifiable result, a deterministic program can perform it for the model.
Alternative title for Zhihu
- Why do large models miscalculate simple math problems?
- How to use tool calls to improve model computing reliability?
- What engineering details do you have to make LLM call calculator?
References and further reading
- Check the OpenAI documentation for your model, SDK and account permissions.
- Python documentation on environment variables, virtual environments and exception handling.
- Consult the documentation for the tools used in the example, such as LangChain, LlamaIndex, Haystack or the vector database.