For Heads-of · Practitioner
Prompt Engineering
The practice of crafting and refining prompts to steer language models towards the outputs wanted, treating the prompt itself as a design problem.
- prompt engineering
- prompting
- technique
- LLM optimisation
Techniques
Chain-of-thought prompting (asking the model to reason step by step), few-shot prompting (giving examples), system prompts (setting context and instructions) and iterative refinement.
Emerging discipline
Small differences in wording can change a model's behaviour significantly, so prompts are increasingly treated as designed, versioned and tested artefacts rather than ad hoc text.
Governance concern
Because phrasing changes results, the same request can produce different outcomes for different users, which raises questions of reproducibility and fairness. A prompt is also not a security boundary: instructions in a system prompt can be overridden by prompt injection or exposed through hidden context exposure. See system prompt hardening.
Related Terms
Context Window
The maximum length of text (measured in tokens) that a language model can consider at once when generating a response.
Hallucination
Output that is fluent, confident and false. It arises from the same mechanism that produces correct output (a model predicting plausible continuations), which is why it is best understood as a property of the technology to be bounded rather than a defect to be fixed.
Large Language Model (LLM)
A deep learning model trained on very large amounts of text, able to generate and interpret human-like language across a wide range of topics and tasks.
Prompt Injection
An attack in which untrusted content reaching a language model is interpreted as instructions rather than data, causing the model to act on an attacker's intent instead of the operator's. It is a property of how the application assembles its context, not a flaw in any particular model.
Related risks and controls
Prompt injection
The attack that prompt wording cannot prevent, however carefully it is engineered.
Hidden context exposure
The engineered prompt is itself an asset, and an extractable one.
System prompt hardening
The part of prompt engineering that is a security discipline rather than a writing one.