AI-Studio/app/MindWork AI Studio/Assistants/PromptOptimizer/prompting_guideline.md
2026-03-24 12:05:00 +01:00

2.6 KiB
Raw Blame History

1 Be Clear and Direct

LLMs respond best to clear, explicit instructions. Being specific about your desired output improves results. If you want high-quality work, ask for it directly rather than expecting the model to guess.

Think of the LLM as a skilled new employee: They do not know your specific workflows yet. The more precisely you explain what you want, the better the result.

Golden Rule: If a colleague would be confused by your prompt without extra context, the LLM will be too.

Less Effective:

Create an analytics dashboard

More Effective:

Create an analytics dashboard. Include relevant features and interactions. Go beyond the basics to create a fully-featured implementation.

2 Add Examples and Context to Improve Performance

Providing examples, context, or the reason behind your instructions helps the model understand your goals.

Less Effective:

NEVER use ellipses

More Effective:

Your response will be read aloud by a text-to-speech engine, so never use ellipses since the engine will not know how to pronounce them.

The model can generalize from the explanation.

3 Use Sequential Steps

When the order of tasks matters, provide instructions as a numbered list.

Example:

1. Analyze the provided text for key themes.
2. Extract the top 5 most frequent terms.
3. Format the output as a table with columns: Term, Frequency, Context.

4 Structure Prompts with Markers

Headings (e.g., # or ###) or backticks (``````) help the model parse complex prompts, especially when mixing instructions, context, and data.

Less Effective:

{text input here}

Summarize the text above as a bullet point list of the most important points.

More Effective:

# Text: 
```{text input here}```

# Task: 
Summarize the text above as a bullet point list of the most important points.

5 Give the LLM a Role

Setting a role in your prompt focuses the LLM's behavior and tone. Even a single sentence makes a difference.

Example:

You are a helpful coding assistant specializing in Python.
You are a senior marketing expert with 10 years of experience in the aerospace industry.

6 Prompt Language

LLMs are primarily trained on English text. They generally perform best with prompts written in English, especially for complex tasks.

  • Recommendation: Write your prompts in English.
  • If needed: You can ask the LLM to respond in your native language (e.g., "Answer in German").
  • Note: This is especially important for smaller models, which may have limited multilingual capabilities.