PALPABLE
VISION


About
Community
Newsletter
Contact

 

Email Me
Book a Call
Privacy Policy
Palpable Vision ©2025

PROMPT INDEX → ENGINEERING

few-shot: Provide a few examples in the prompt to guide the model’s response style or structure.

zero-shot: Ask the model to complete a task without giving examples, relying solely on instructions.

chain-of-thought: Prompting the model to reason step-by-step before producing a final answer.

tree of thoughts: Explore multiple reasoning paths in parallel to evaluate and choose the best outcome.

meta prompting: Ask the model to generate or improve prompts themselves, optimizing task performance.

RAG (retrieval augmented generation): Combining external data retrieval with the model’s generation to provide grounded, updated responses.

schema prompting: Designing prompts to produce structured outputs (JSON, tables) that align with predefined schemas.

modular prompting: Organizing prompts into reusable, logical blocks for clarity, scalability, and control.

ReAct prompting: Combining reasoning steps with actions (tool use or API calls) in a single workflow.

reflective prompting: Instruct the model to critique, revise, or improve previous responses before finalizing it.


Prompt Structure:
Model → Choose mental model
Role → who should the AI be / viewpoint
Problem → description of the situation / challenge
Goal → what success looks like
Instructions → steps the model must follow
Context → audience, constraints, inspiration

Example:

Model: First Principles Thinking
Role: Productivity Analyst (10+ years’ in habit and creative workflow design).
Problem: Rarely finish <side project>, stalling after week one.
Goal: Design a finish‑without‑burnout system.
Instructions:
1. List the core factors behind finishing
2. Expose one “obvious” assumption per factor
3. Rewrite factors into a lean completion framework
4. Spot one weakness in the new system + a quick safeguard.
Context: audience = solo creatives, 5–7 hrs/week free