The Principle Pyramid
```text LEVEL 1: UNIVERSAL FOUNDATIONS (Work across ALL AI systems)
- Clear Communication - Specific > Vague - Concrete > Abstract - Simple > Complex
- Logical Structure - Sequential thinking - Hierarchical organization - Cause-effect clarity
- Goal Orientation - Define success - Specify outputs - Measure results
LEVEL 2: PATTERN PRINCIPLES (Adapt to model categories)
- Instruction Models - Command clarity - Role definition - Task decomposition
- Chat Models - Conversation flow - Context retention - Personality consistency
- Creative Models - Inspiration seeding - Style guidance - Iteration cycles
LEVEL 3: MODEL SPECIFICS (Fine-tune for each system)
- Syntax preferences - Optimal prompt length - Special features - Known quirks ```text
Cross-Model Translation
```text PROMPT TRANSLATION MATRIX:
Original (GPT-4): "You are an expert analyst. Examine this data and provide three key insights with supporting evidence."
Claude Translation: "I need your analytical expertise. Please examine this data carefully and identify three significant insights. For each insight, provide specific supporting evidence from the data."
Gemini Translation: "Analyze the following data from an expert perspective. Identify and explain three key insights, citing specific evidence for each."
Llama Translation: "### Task: Expert Data Analysis Examine the provided data and identify three key insights. For each insight, include: - The finding - Supporting evidence - Implications"
TRANSLATION PRINCIPLES: - Maintain core intent - Adapt communication style - Leverage model strengths - Avoid model weaknesses ```text