Glossary

AI in plain English.

Definitions for the terms you'll see across our prompts and guides.

C

Context window: The maximum amount of text (in tokens) a model can consider at once. Larger windows let you feed in more documents or longer conversations.
Chain of thought: Instructing the model to reason step-by-step in writing before producing a final answer. Often improves accuracy on multi-step problems.
Code-switching: Mixing languages within a sentence or paragraph, common in Singapore (Singlish), Malaysia (Manglish), Philippines (Taglish), and many other multilingual contexts.

Embedding: A vector representation of text used to find semantically similar content. Powers most retrieval and search features.

Few-shot prompting: Including a small number of input/output examples in the prompt to teach the model the pattern you want.
Fine-tuning: Training a base model on extra examples to specialise it for a domain or style. Usually overkill compared to good prompting and RAG.

Hallucination: When a model generates content that sounds plausible but is factually wrong. Mitigated with retrieval (grounding the model in real documents) and careful prompt design.
Honorifics (Korean): Korean speech levels (-습니다, -해요, -해, etc.) that mark formality and the speaker-listener relationship. Vital in business communication.

Keigo (敬語): Japanese honorific language. Three layers — sonkeigo (respectful), kenjougo (humble), teineigo (polite) — applied based on relationships and context.

Prompt: The instruction or question you send to an AI model. Better prompts produce better outputs; the craft of writing them is sometimes called prompt engineering.

RAG (Retrieval-Augmented Generation): Architecture that retrieves relevant documents at query time and includes them in the prompt, so the model answers from your data.

System prompt: An instruction that frames the AI's role, persona, or rules across the whole conversation, separately from the user's individual messages.

Token: A unit of text (roughly 3-4 characters in English, often a single character in CJK languages) that AI models read and generate. Pricing and context limits are usually denominated in tokens.
Temperature: A model parameter controlling randomness. Lower = more deterministic, higher = more creative/variable.