AI Common Terms
0:000:00
Common AI-related terms
✅ AI Terms & Their Simple Meanings
| Term | Simple Meaning |
|---|---|
| LLM (Large Language Model) | A big AI model trained on lots of text that can understand and generate language |
| Fine-tuning | AI is retrained using your data to improve its knowledge or behavior permanently |
| Grounded AI | Answers only from your data |
| Retrieval | Finds extra info from a document, database, or internet when needed |
| Live Web Access | Searches the internet in real time to answer questions |
| Knowledge Cutoff | The latest date the AI was trained on — it doesn't know anything after that |
| RAG (Retrieval-Augmented Generation) | AI looks up info from your data before answering |
| Embedding | Converts your data into a special format (vectors) so AI can search and match better |
| Prompt Engineering | Designing smart questions or instructions to get better answers from AI |
| Inference | The act of the AI answering a question or doing a task (based on its training/data) |
| Hallucination | When AI makes up a wrong or fake answer confidently |
| Context Window | The memory space AI uses at once — like how many words it can "see" at a time |
| Token | A piece of a word the AI understands — like a puzzle piece for sentences |
| Chat Memory | What the AI remembers in a conversation (temporary unless stored intentionally) |
| Agents | AI tools that take actions or make decisions on their own using rules or goals |