Skip to content
LunaNotes home

Generative AI Career: Top 10 Basic Keywords Explained Simply

Convert to note

Why Focus on the Basics?

To build a career in AI, you need strong fundamentals. This series starts from the basics and moves to advanced concepts. The video covers ten common Generative AI keywords to get you started. For a more comprehensive overview, consider exploring Master Generative AI: From Basics to Advanced LangChain Applications.

Top 10 Generative AI Keywords Explained

1. Token and Tokenization

  • What is a Token? A token is the smallest building block of text. A sentence is split into smaller parts (words or subwords) called tokens.
    • Example: 'Hello I am a token' becomes 5 tokens (including spaces).
  • What is Tokenization? The process of converting text into tokens. A tokenizer tool (like from OpenAI) shows this conversion.
  • Byte Pair Encoding (BPE): The algorithm most models use for tokenization.

2. Large Language Model (LLM)

  • What is an LLM? A model trained on massive datasets. Better data quality and quantity lead to better model performance (e.g., GPT-4, GPT-5).
  • How it works: When you ask ChatGPT a question, it first tokenizes your text, then processes those tokens.

3. Context Window

  • Definition: The maximum number of tokens an LLM can process and generate in one go.
  • Example: A small model might have a 512-token context window. If your PDF is 1000 tokens, the model can only process half, truncating the rest and losing context.

4. Chunking

  • Problem: You cannot send a whole large document (like a 1000-line PDF) to an LLM at once. It would exceed the context window.
  • What is Chunking? A process that splits large documents into smaller, manageable pieces (chunks) while preserving meaning.
  • Chunking Strategy: The algorithm you use to split text matters. Otherwise, chunks can lose context and meaning.

5. Embeddings

  • What are Embeddings? When you create chunks, you convert them into numerical vectors.
  • Example: Words like 'King', 'Queen', and 'Royal' have similar vectors because they belong to the same category.
  • Use: These vectors are stored in a Vector Database. When you ask a question, the system finds similar vectors to retrieve relevant chunks.

6. Vector Database

  • Role: Stores the numerical vectors (embeddings) created from your chunks.
  • Process: Based on your query's context, it retrieves the most relevant vectors and sends those chunks to the LLM.

7. Transformers

  • Definition: The architecture on which most LLMs are built. It has two main components:
    1. Encoder: Converts your text prompt into numerical representations (tokens/numbers).
    2. Decoder: Works on those numbers to generate the final response.
  • Key Components: Encoder-Decoder model, Feed Forward networks, and Attention Mechanism. For a deeper dive into how these models work, read Understanding Generative AI: Concepts, Models, and Applications.

8. Temperature

  • Definition: A hyperparameter that controls the randomness and creativity of an LLM's output.
  • Range & Effect:
    • High Temperature (e.g., 1.0): Very creative, imaginative, or fictional responses.
    • Low Temperature (e.g., 0.0): Precise, factual, and deterministic answers.
  • Note: Other factors like Top-K and Top-P also influence output.

9. RAG (Retrieval-Augmented Generation)

  • Problem: LLMs only know public data. They cannot answer from your private, confidential company data without being trained on it (which is expensive).
  • How RAG Works:
    1. Retrieve: Take your user's question, find the most relevant chunks from your private documents (using embeddings and vector DB).
    2. Augment: Combine the retrieved chunks with the user's original prompt.
    3. Generate: Send this combined context to the LLM, which then generates an answer based on that specific data.
  • Keywords Used: Tokenization, Chunking, Embeddings, and Vector Databases are all part of the RAG pipeline. To understand this architecture in more detail, check out Understanding Retrieval Augmented Generation (RAG) in AI Applications.

10. Hallucination

  • Definition: When an LLM gives an incorrect answer with high confidence.
  • Control Factors:
    • Temperature: Lower temperature reduces hallucination.
    • Data Quality: Your chunks and embeddings must be accurate.
    • Chunking Strategy: Poor chunking can lead to bad context and wrong answers.
    • Embedding Model: The model must match the type of data (e.g., don't use a text embedding model for chart data).

Summary: How Keywords Connect in a RAG Workflow

  1. Data Ingestion Phase: Data is split into chunks, then tokenized, and finally converted into embeddings for storage in a Vector Database.
  2. Retrieval & Assembly Phase: The Vector Database finds the nearest neighbors (similar vectors) based on context. RAG injects the top chunks into the prompt, packaging them according to the Context Window.
  3. Inference & Output Phase: Transformers (self-attention) process the data, Temperature controls the output style, and good data reduces Hallucination. For a broader perspective on the landscape, see Understanding Generative AI, AI Agents, and Agentic AI: Key Differences Explained.

Final Advice: Focus on fundamentals first. Don't jump into complex diagrams. Be patient, take notes, stay curious, and you will progress from basics to production without frustration. To add to your vocabulary, review 20 Key AI Terms Every Engineer Must Know Explained Simply.

Keep this summary

Save it to LunaNotes and it becomes a real note in your library — editable, searchable, and ready to turn into flashcards or a diagram. Free to start.

Save to LunaNotes

Or summarise for another video.

This summary and transcript were automatically generated using AI with the Free YouTube Transcript Summary Tool by LunaNotes.

Related summaries

Understanding Retrieval Augmented Generation (RAG) in AI Applications

Understanding Retrieval Augmented Generation (RAG) in AI Applications

Discover how Retrieval Augmented Generation (RAG) enhances large language models by integrating external knowledge bases to overcome hallucination and improve accuracy. This comprehensive overview explains RAG's workflow, its advantages over traditional LLM approaches, and upcoming implementation strategies using vector databases and embeddings.

Master Generative AI: From Basics to Advanced LangChain Applications

Master Generative AI: From Basics to Advanced LangChain Applications

Explore the comprehensive journey into generative AI, from foundational concepts and transformer architectures to practical implementation with LangChain. Learn how to leverage large language models, prompt engineering, retrieval augmented generation, and ChatGPT-like systems to build cutting-edge AI applications and stay ahead in the evolving AI landscape.

Understanding Generative AI: Concepts, Models, and Applications

Understanding Generative AI: Concepts, Models, and Applications

Explore the fundamentals of generative AI, its models, and real-world applications in this comprehensive guide.

20 Key AI Terms Every Engineer Must Know Explained Simply

20 Key AI Terms Every Engineer Must Know Explained Simply

This video breaks down 20 essential AI and large language model (LLM) terms, from tokenization and vectors to RAG and reinforcement learning. Designed for engineers, it clarifies confusing jargon and connects concepts like transformers, fine-tuning, and MCP to real-world applications.

Understanding Generative AI, AI Agents, and Agentic AI: Key Differences Explained

Understanding Generative AI, AI Agents, and Agentic AI: Key Differences Explained

In this video, Krishna breaks down the essential differences between generative AI, AI agents, and agentic AI. He explains how large language models and image models function, the role of prompts in generative applications, and the collaborative nature of agentic AI systems.

Most viewed summaries

A Comprehensive Guide to Using Stable Diffusion Forge UI

A Comprehensive Guide to Using Stable Diffusion Forge UI

Explore the Stable Diffusion Forge UI, customizable settings, models, and more to enhance your image generation experience.

Kolonyalismo at Imperyalismo: Ang Kasaysayan ng Pagsakop sa Pilipinas

Kolonyalismo at Imperyalismo: Ang Kasaysayan ng Pagsakop sa Pilipinas

Tuklasin ang kasaysayan ng kolonyalismo at imperyalismo sa Pilipinas sa pamamagitan ni Ferdinand Magellan.

Mastering Inpainting with Stable Diffusion: Fix Mistakes and Enhance Your Images

Mastering Inpainting with Stable Diffusion: Fix Mistakes and Enhance Your Images

Learn to fix mistakes and enhance images with Stable Diffusion's inpainting features effectively.

Pamamaraan at Patakarang Kolonyal ng mga Espanyol sa Pilipinas

Pamamaraan at Patakarang Kolonyal ng mga Espanyol sa Pilipinas

Tuklasin ang mga pamamaraan at patakaran ng mga Espanyol sa Pilipinas, at ang epekto nito sa mga Pilipino.

How to Install and Configure Forge: A New Stable Diffusion Web UI

How to Install and Configure Forge: A New Stable Diffusion Web UI

Learn to install and configure the new Forge web UI for Stable Diffusion, with tips on models and settings.

Found this summary useful?

Take it with you. One click puts it in your own LunaNotes library.

Save to LunaNotes

Start taking better notes today with LunaNotes

Start for free