Prompting
Context compression
Context compression shortens input while preserving key information.
Quick definition
Context compression shortens input while preserving key information.
- Category: Prompting
- Focus: control and response quality
- Used in: Standardizing responses with templates and clear instructions.
What it means
Summaries or extraction reduce token usage. In prompting workflows, context compression 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 a long thread before continuing.
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.
