The AI-900 exam has been retired
It was retired on June 30, 2026. It was replaced by AI-901 (Microsoft Azure AI Fundamentals). Go to the AI-901 study guide →
6.1.2. Model Parameters and Prompt Engineering
Understanding how to control model behavior is essential for effective generative AI applications.
Tokens and Context Windows: LLMs process text in chunks called tokens (roughly 4 characters or 0.75 words in English). The context window is the maximum number of tokens the model can consider at once. Larger context windows enable:
- Longer conversations
- Processing larger documents
- More context for better responses
Temperature and Sampling: Temperature controls response randomness:
- Low temperature (0.0-0.3): More deterministic, focused responses
- High temperature (0.7-1.0): More creative, diverse responses
⚠️ Exam Trap: DALL-E generates images but CANNOT analyze or describe images. Image description requires a vision-capable model like GPT-4 with vision. DALL-E's capabilities are: creating new images, creating variations, and editing images.
Prompt Engineering: Effective prompts significantly impact response quality:
| Technique | What It Does | Example |
|---|---|---|
| Zero-shot | Ask directly without examples | "Translate to French: Hello" |
| Few-shot | Provide examples first | "dog→chien, cat→chat, hello→?" |
| Chain-of-thought | Ask for step-by-step reasoning | "Think through this step by step..." |
| System messages | Set context and constraints | "You are a helpful assistant that..." |
