Prompting
Prompt
A prompt is the input text that guides the model's response.
Quick definition
A prompt is the input text that guides the model's response.
- Category: Prompting
- Focus: control and response quality
- Used in: Standardizing responses with templates and clear instructions.
What it means
Prompts can include instructions, context, and examples. In prompting workflows, prompt often shapes control and response quality.
How it works
Prompting structures the instructions, context, and examples the model sees. Small changes in wording and format can shift output quality.
Why it matters
Prompting controls behavior, format, and response quality.
Common use cases
- Standardizing responses with templates and clear instructions.
- Improving quality with few-shot examples.
- Constraining format for JSON, tables, or checklists.
Example
Summarize this article in 5 bullets.
Pitfalls and tips
Overly long prompts dilute key instructions and waste tokens. Keep prompts concise and explicit.
In BoltAI
In BoltAI, you will see this in prompts, system instructions, and templates.
