Prompt Efficiencies

Posted on: 15 August 2025

Tokens

A token is a small chunk of text that the AI model uses to process your message and generate a reply. Both what you enter and what the model sends back count toward your token limit. When you write short, focused prompts and ask for concise answers, you use fewer tokens overall. That means your sessions can last longer, responses come faster, and you’re less likely to hit token limits or get cut off mid-reply.

Conserve tokens = longer sessions + faster responses + fewer cutoffs

1️. Prompt Design

Do

• Keep prompts concise: use direct verbs (“summarize,” “revise,” “explain briefly”).
• Reference earlier text instead of re-pasting it.
• State the output size:
  • “In 3 bullet points.”
  • “Under 100 words.”
  • “Code only—no explanation.”

Don’t

• Repeat long setup text (“As I mentioned earlier…”).
• Ask for multiple tasks in one prompt if they can be done sequentially.
• Use filler (“Could you please kindly explain…”).

2. Conversation Management

• Stay in the same chat thread so context persists.
• Split long projects:
  1. Outline
  2. Draft one part
  3. Revise iteratively
• Delete or summarize unneeded sections before continuing.

3️. Handling Large Text

• Paste only relevant sections (“Here’s the last 30 lines”).
• Summarize before submission (“This 20-page doc covers X, Y, Z…”).
• Use compressed summaries or keywords when possible.

4. LLM-Specific Tips

5️. Output Control

• Ask for structured formats (Markdown, table, JSON).
• Set maximums:
  • “Explain in ≤5 sentences.”
  • “Show first 10 lines only.”
• For code: “Return only the function definition.”

Approximate Token Use

More posts:

• Next: LLM Prompt Organizer