LibGuides: Scholarly Use of AI tools: Prompt Engineering

What is Prompt Engineering?

The process of crafting and refining natural-language inputs to guide GenAI tools and models, particularly large language models (LLMs), to produce desired outputs. Effectively designing such conversational inputs helps these AI systems understand the user's intent, follow instructions, and generate relevant and accurate responses.

6 Key Principles for Effective Prompt Design

1. Keep It Simple

Use clear, straightforward language.
Avoid unnecessary complexity.
Start simple: Begin with a basic prompt and add details only if needed.

2. Be Specific

Include the necessary level of detail to achieve the desired result.
Avoid ambiguous or vague expressions.
Use concrete terms instead of general or unclear wording.

3. Provide Context

Give relevant background information
State your purpose
Include any necessary constraints

4. Structure Complex Requests

Break large tasks into smaller steps
Create a sequence of focused prompts
Build on previous responses

5. Define Your Output

Specify the desired format (poetry, list, code, etc.)
Set clear parameters (length, style, tone, etc.)
Indicate level of detail needed

6. Iterate & Refine

Test and adjust your prompts
Refine prompts based on past responses
Try different approaches if needed

Quick Tips:

Modularly combine these principles as required.
Start with a basic prompt and refine it based on the response.
Simplicity and specificity work in tandem: your prompts should be as simple as possible while remaining as specific as necessary.
Remember that effective prompting is an iterative process.