Explore automated prompt engineering techniques using DSPy in this 46-minute video tutorial. Learn how to leverage structured prompt generation for improved AI model performance. Dive into topics like simple prompting, chain of thought, vector search retrieval, few-shot examples, multi-hop queries, and result optimization. Gain hands-on experience with practical notebooks, installation guidance, and step-by-step demonstrations. Discover how to enhance answer quality control using DSPy assertions and compare overall results across different approaches. Access additional resources, including slides, notebooks, and timestamps for easy navigation through the content.
Automated Prompt Engineering with DSPy - A Comprehensive Tutorial
Overview
Syllabus
Introduction to structured/automated prompt generation
Video Overview
Why is DSPy useful?
Q&A and Retrieval Databases hotpotqa and wikipedia
Notebooks - DSPy and Trelis
Installation and Setup
Step 1: Simple Prompting and Benchmarking
Step 2: Adding Chain of Thought
Step 3: Add vector search retrieval
Step 4: Add random few shot examples
Step 5: Optimally choosing few shot examples
Step 6: Add multi-hop search/queries Baleen / Perplexity-style
Step 7: Multi-hop search WITH optimally chosen few shot examples
Overall Results Comparison!
DSPy assertions - for further answer quality control
Video resources
Taught by
Trelis Research
Related Courses
-
Optimizing LLM Pipelines with DsPy - Part 2
-
DSPy Tutorial - Programming Language Models with 8 Key Examples
-
The Prompt Report: Understanding Prompting Techniques and Taxonomy - Part 1
-
AI Mastery: Ultimate Crash Course in Prompt Engineering for Large Language Models
-
Prompt Engineering Techniques for Microsoft Phi3-mini 128k Instruct
-
Prompt Engineering Techniques with Mistral 7B Instruct for Data Science and Machine Learning
