Introducing a GitHub repository full of hot tips and instructions on how to build the perfect prompt presented in a collection of Jupiter Notebooks.

While the whole idea of prompting instead of directly asking a question might sound silly, it does actually play a big role. For instance if you use a prompt like "You are an expert in computer science. Tell me how a CPU works" rather than just asking "Tell me how a CPU works" you'll find that the first (prompt) will bring back far superior results than asking directly and without the embellishment. This is because the former has set the context for the AI to work on.

Of course, the more elaborate the prompt the better the outcome. But how do I construct a really great prompt?
The answer to that question is given by this "Prompt Engineering Techniques " Github repository which contains the whole know-how.

Its target group encompasses beginners getting their feet wet through to seasoned GenAI practitioners who push the boundaries of what's possible. That said, it doesn't just offer a mere exploration of techniques, but it showcases too how to use those techniques from within code, in Python of course, calling the Langchain library. 

The structure is laid out as a course would be. You start with the fundamentals and increasingly move on to more complex techniques. A quick overview follows.

The fundamental chapters offer a comprehensive introduction to the fundamental concepts of prompt engineering in the context of AI and language models, and explore the three core techniques of Zero-Shot Prompting, Few-Shot Learning and Chain of Thought (CoT) Prompting.

The chapter on Advanced Strategies explores techniques for generating diverse reasoning paths and aggregating results to improve AI-generated answers and focuses on techniques to set up constraints for model outputs and implement rule-based generation.

Advanced Implementations explores techniques for breaking down complex tasks and chaining subtasks in prompts or connect multiple prompts, followed by Prompt Optimization Techniques and Handling Ambiguity and Improving Clarity.

There's also an exploration of specialized techniques like negative prompting and techniques for avoiding undesired outputs from large language models, prompt formatting and the creation and use of prompts for specific tasks such as text summarization, question-answering, code generation, and creative writing.

Finally the course ends up with taking a look into some advanced applications :

• Multilingual and Cross-lingual Prompting
• Prompt Security and Safety
• Evaluating Prompt Effectiveness

Everything is bundled up together in a single Github repository as a collection of Jupiter Notebooks.

To sum it up, Prompt Engineering Techniques really has everything covered, at the same time managing to address both beginners and intermediates upgrading them to experts. The author, Nir Diamant, maintains two other great repositories for generative AI as well, "RAG techniques" and "GenAI Agents". Make sure to check them out.

More Information

Prompt Engineering Techniques 

Related Articles

Microsoft's Generative AI for Beginners Version 2

To be informed about new articles on I Programmer, sign up for our weekly newsletter, subscribe to the RSS feed and follow us on Twitter, Facebook or Linkedin.

Comments

Make a Comment or View Existing Comments Using Disqus

or email your comment to: comments@i-programmer.info