LLM Hacks v1.0 - Enhancing Reasoning, Response Quality, and Development

Release Title:

LLM Hacks v1.0 - Enhancing Reasoning, Response Quality, and Development

Release Notes:

This release includes a collection of prompts and techniques aimed at improving the capabilities of Large Language Models (LLMs), focusing on reasoning, response development, and quality enhancement.

Key Features:

• Deductive Reasoning Enhancer:

  • Improves LLM's deductive reasoning capabilities.
  • Hidden reasoning steps for streamlined interaction.

• Response Development Enhancer:

  • Guides the LLM through structured problem-solving.
  • Encourages consideration of multiple perspectives.

• Response Quality Enhancer:

  • Token-efficient prompt focused on improving final response quality.
  • Includes a self-evaluation checklist.

• Response Quality Enhancer Plus:

  • Builds on the basic quality enhancer with Chain-of-Thought (CoT) principles.
  • Ensures logical consistency and comprehensive answers.

• Merged Best Practice Prompt:

  • Combines all enhancers in a unified approach for comprehensive LLM interaction improvements.

Key Files:

• merged_best_practice_prompt.txt: A single prompt combining deductive reasoning, response development, and quality enhancement.

Usage:

Include these prompts at the start of your LLM interactions or in custom instructions. Detailed instructions are available in each sub-directory.

Acknowledgments:

• Based on work from Awesome_GPT_Super_Prompting, "On the Diagram of Thought", and "Chain of Thought Empowers Transformers".

Feedback and contributions are welcome.