Intermediate7 prompts

Chain of Thought

Chain-of-thought (CoT) prompting instructs the model to reason step by step before answering. By working through intermediate reasoning stages explicitly, the model achieves significantly higher accuracy on math, logic, and multi-step problems — and its answers become auditable.

How It Works

1

Ask the model to "think step by step" or "show your reasoning".

2

The model decomposes the problem into sequential logical steps.

3

Each step builds on the previous one, reducing reasoning errors.

4

The final answer follows the visible reasoning trail.

Real Example

Prompt: "A store gives 20% off an $80 jacket and then adds 8% tax. What is the final price? Think step by step." The model computes: discount = $16, price = $64, tax = $5.12, final = $69.12.

Best Practices

  • Explicitly request step-by-step reasoning.
  • Ask the model to verify its own conclusion after reasoning.
  • For math, request the formula before the number.
  • Combine CoT with few-shot examples for hard problems.

Common Mistakes

  • Assuming the model reasons when you do not ask it to.
  • Forcing unnecessarily long reasoning for trivial tasks.
  • Not cross-checking the final numeric answer.

When to Use CoT

Math and logic problemsMulti-step decisionsDebugging and root-cause analysis

Frequently Asked Questions

When should I use Chain of Thought?

Math and logic problems, Multi-step decisions, Debugging and root-cause analysis. If your task is more complex or your output is inconsistent, consider combining it with examples or chain-of-thought.

Does Chain of Thought work on all AI models?

Most prompting techniques transfer across modern models like ChatGPT, Claude, and Gemini, though results vary. Test the same prompt on two models and keep the best performer.

What is the next technique to learn after CoT?

Based on your related techniques, explore: Few Shot Prompting, Tree of Thoughts, ReAct Prompting.

How do I practice CoT?

Use the prompts above as starting points, then modify one variable at a time. Track which changes improve your results in a prompt library.

Skip to content