GENERATIVE-AI-LEADER · Question #80
A creative team at example.com is using a large language model to craft ad taglines and notices that asking "Create a tagline" returns bland ideas. When they instead ask "Write a punchy and…
The correct answer is C. Prompt engineering. Prompt engineering is the practice of deliberately crafting and refining the input given to a language model-specifying context, tone, constraints, audience, and detail-to steer the model toward higher-quality, more relevant outputs. The example shows moving from a vague prompt…
Question
A creative team at example.com is using a large language model to craft ad taglines and notices that asking "Create a tagline" returns bland ideas. When they instead ask "Write a punchy and memorable tagline for a new fair trade matcha tea subscription that highlights plastic free packaging and a smooth calm energy, aimed at remote workers in major cities ages 22 to 32," the outputs are far more relevant and engaging. What is the practice of deliberately shaping the input to the model to obtain better results called?
Options
- AReinforcement learning from human feedback or RLHF
- BModel fine-tuning
- CPrompt engineering
- DData augmentation
How the community answered
(24 responses)- B4% (1)
- C96% (23)
Explanation
Prompt engineering is the practice of deliberately crafting and refining the input given to a language model-specifying context, tone, constraints, audience, and detail-to steer the model toward higher-quality, more relevant outputs. The example shows moving from a vague prompt to a richly specified one, which is the core activity of prompt engineering. RLHF (A) is a training technique using human feedback signals. Fine-tuning (B) re-trains model weights on domain-specific data. Data augmentation (D) expands training datasets, not prompts.
Topics
Community Discussion
No community discussion yet for this question.