Domain: Prompting & Instructions

Explicit output constraints (format, tone).

Breaking tasks into sub-steps.

Using markers to isolate context segments.

Multiple examples included in prompt.

One example included to guide output.

The text (and possibly other modalities) given to an LLM to condition its output behavior.

Crafting prompts to elicit desired behavior, often using role, structure, constraints, and examples.

Asking model to review and improve output.

Prompt augmented with retrieved documents.

Assigning a role or identity to the model.

Temporary reasoning space (often hidden).

Sampling multiple outputs and selecting consensus.

Enables external computation or lookup.

Task instruction without examples.