- prompt -> LLM -> output
- How efficiently you provided the prompt, you get the output of the LLM that efficient.
- The quality of an LLM’s output depends on how well you craft the prompt.
| Prompt Type | Description | Example |
|---|---|---|
| Zero-shot prompting | No examples, just ask directly | "Translate this into German: 'Good morning'" |
| Few-shot prompting | Provide a few examples to guide the model | "'Hello → Hallo', 'Thanks → Danke', 'Goodbye → ?'" |
| Chain-of-thought | Ask the model to reason step by step | "What’s 15 + 27? Let’s think step by step." |
| Instruction-based | Clear task description in natural language | "Summarize this text in one sentence." |
- Good prompts significantly reduce the cost for the organisations.
- Good and bad prompts affect cost in LLMs because clearer, more efficient prompts reduce the number of tokens and retries needed, directly lowering usage and compute expenses.
(1*) DAY-1 | Fundamentals of AI Assisted DevOps | Demo and Notes Included