Prompt Engineering Strategies from Popular AI Assistants
Analyzing system prompts from popular AI assistants to extract common prompt engineering strategies.
Using the awesome-ai-system-prompts repository as source data, which contains leaked system prompts from major AI systems like ChatGPT, Claude, Manus, and others, we've used DocETL to analyze and extract structured information about strategies employed across these different AI assistants.
Our pipeline identifies:
- Common strategies: Recurring patterns and approaches used across different AI systems
- Strategy implementation examples: How specific AI systems implement these strategies
- Strategy summaries: Concise explanations of each strategy and its implementation considerations
This analysis can help prompt engineers and AI developers identify best practices and effective approaches when designing system prompts for their own applications.
The entire pipeline processing cost just $0.18 to run for all 19 systems.
Note: The strategies identified are extracted through automated analysis and may not capture all nuances or context. Use these insights as a starting point for your own prompt engineering work.