CloudYetiFREE
AI tip
Claude follows a rule better when you say why
A rule in capital letters with no reason is the weak version. Add the reason after "so" or "because", and Claude can handle the cases you did not list.
Why it matters
Most rules people give an AI look like "NEVER use ellipses": a word in capitals and no reason.
Anthropic's prompting guide shows the stronger version: the same rule with the reason in the sentence.
The guide's words: "Claude can generalize from the explanation."
Do this
- Open your system prompt, project instructions or CLAUDE.md.
- Find every rule with NEVER, ALWAYS or MUST in it.
- Add the reason after the word "so" or "because".
- Run the same task again and compare the answer.
Anthropic's example of the stronger rule
Your response will be read aloud by a text-to-speech engine, so never use ellipses since the text-to-speech engine will not know how to pronounce them.
One thing to keep in mind: The reason has to be the real one. A made-up reason teaches Claude the wrong goal, and it will generalize from that too.
The less effective and more effective rules are Anthropic's own example from the guide. Source: Anthropic: Prompting best practices, "Add context to improve performance".
Ask me anything about AI and cloudA paid call on your own setup: your agents, your bill, your plan.Book a call →