prompting

_{scroll ↓ to Resources}

Contents

• Note
  • [[#Note#Examples|Examples]]
  • [[#Note#Reasoning|Reasoning]]
  • [[#Note#Other|Other]]
• Resources

Note

Extensive list of prompting methods from ^44dbf0

Examples

• zero-shot learning
• few-shot learning
  • append sample task-answer examples to the prompt
  • Many-Shot In-Context Learning shows how the results are impacted if the prompt is filled with 2000+ examples. Though, consider the costs of such prompting technique and difficulty of generating a prompt itself.

Reasoning

• Chain-of-Thought - add let’s think step by step to your prompt
  • Chain-of-Thought Prompting Elicits Reasoning in Large Language Models
• self-consistency + CoT - generate several different LLM outputs to the same prompt, take the most often answer
• Tree-of-Thought - iteratively generate different candidate outputs, pick the best and add them back to the tree of potential solutions
  • Tree of Thoughts: Deliberate Problem Solving with Large Language Models
• Graph-of-Thought - Graph of Thoughts: Solving Elaborate Problems with Large Language Models
• Reasoning with Large Language Models, a Survey

Other

• meta prompting
  • ask the model to improve own answer

Resources

• An Empirical Categorization of Prompting Techniques for Large Language Models: A Practitioner’s Guide
• A Systematic Survey of Prompt Engineering in Large Language Models: Techniques and Applications

---

Links to this File

table file.inlinks, file.outlinks from [[]] and !outgoing([[]])  AND -"Changelog"