Prompt -The instruction or question you give an AI.
Token – The smallest unit of text an AI model processes (roughly three-quarters of a word).
LLM (Large Language Model) – The type of AI behind ChatGPT, Claude, Gemini, Llama
Hallucination – When an AI confidently gives false or invented information.
Temperature – A setting (0'1 or 0'2) that controls how creative vs. predictable the AI is. Low = factual and focused; High = more imaginative (good for brainstorming).
Fine-tuning – Training an AI on your specific data or style so it sounds like you or your company. Companies now fine-tune models on internal docs for brand voice. |
Agent – An AI that can take actions on its own (book appointments, send emails, run code).
Multimodal – AI that handles text + images + audio + video (GPT-4, Gemini 1.5, Claude 3). Upload a photo of a skin rash or a spreadsheet and get analysis.
Context window – How much text the AI can remember in one conversation (128k or 1M tokens).
Chain-of-thought – Telling the AI to, think step by step' to get better reasoning. Dramatically improves math, logic, and planning.
Guardrails – Rules built into an AI to prevent harmful, biased, or off-brand outputs.
Inference – The actual running of the AI to generate answers (training it).
Mode – The actual AI brain (GPT-4, Claude 3.5 Sonnet, Gemini 1.5 Pro, Grok-2). People now compare AIs the same way we compare car engines.
