Prompt Engineering

The practice of carefully designing the text prompt to get better outputs from a language model. Includes choosing the number of examples, the format (bullet points, line breaks, etc.), the phrasing of instructions, and adding role instructions like "You are a sentiment classifier."

The practice of carefully crafting input prompts to achieve desired model behavior without changing model weights. CoT is fundamentally a prompt engineering technique — no model retraining required. The quality of CoT examples (clarity, step size, coverage) significantly affects performance, making prompt engineering critical.