Methods for effective communication with large language models
Core Idea: Prompt engineering encompasses methods for communicating with large language models to guide their behavior toward desired outcomes without modifying model weights, relying on empirical testing and heuristics to discover effective prompting patterns.
Key Elements
Core Techniques
- Chain-of-Thought (CoT) Prompting
- Few-Shot Prompting
- Zero-Shot Prompting
- Instruction Prompting
- Chain Prompting
- Tree of Thoughts (ToT)
- Emotional Prompting with AI
Frameworks and Methodologies
- Prompt Development Frameworks
- Prompt Engineering Principles
- Automatic Prompt Design
- Effective AI Prompting Techniques
Application-Specific Approaches
Tools and Evaluation
- Prompt Testing Tools
- Prompt Engineering Tools
- Prompt Engineering Tools for Agentic Systems
- Prompt Visualization Tools
- Prompt Engineering Frameworks
Additional Connections
Theoretical Foundations
- In-context Learning
- Few-shot Learning
- Instruction Tuning
- RLHF (Reinforcement Learning from Human Feedback)
Broader Applications
Related Concepts
References
-
Primary Source:
- Weng, Lilian. (Mar 2023). Prompt Engineering. Lil'Log. https://lilianweng.github.io/posts/2023-03-15-prompt-engineering/
-
Additional Resources:
- White, Jules et al. (2023). A Prompt Pattern Catalog to Enhance Prompt Engineering with ChatGPT. arXiv preprint arXiv:2302.11382.
- Reynolds, Laurie & McDonell, Kyle. (2021). Prompt Programming for Large Language Models: Beyond the Few-Shot Paradigm. CHI Extended Abstracts.
- Anthropic. (2023). Claude Prompt Design Documentation. https://www.anthropic.com/index/claude-prompt-design
- OpenAI. (2023). OpenAI Cookbook: Prompt Engineering Guide. https://github.com/openai/openai-cookbook
- LangChain Documentation. (2023). Prompt Templates. https://docs.langchain.com/docs/components/prompts/prompt-templates
#prompt-engineering #LLM #NLP #machine-learning #AI-alignment #model-steerability #language-models #AI-interaction #context-learning #prompt-patterns
Sources: