|
This month we are looking at Google’s recent iteration on their Prompt Engineering guide as a vehicle to discuss the latest advancements in the field of prompting models. We will cover the core material of the paper and discuss the sort of the things that are left out of the current guidance, and potential resources to deepen understanding of different prompt engineering components.
The focus of the discussion will prioritize programmatic control first, however many concepts will also be useful for general users a well.
Topics for the session will primarily focus on:
- Core components of a prompt
- Key Prompting Parameters (Temperature, Top-p, Top-k, etc.)
- Prompting Techniques (Few-Shot, Chain of Thought, Self-Consistency, Tree of Thoughts, ReAct, etc.)
- Prompting Modes (Structured Outputs, Vision, Tool Use, Reasoning )
- Prompt Evaluation & Automating Prompt Engineering (DSPy, etc)
|
