Lesson 3 of 10
Lesson 3 — Learning Hub

ChatGPT & Large Language Models – How They Work and How to Use Them Well

11 min read
Beginner

What Are Large Language Models?

A Large Language Model, or LLM, is a type of deep neural network trained on massive amounts of text — billions of web pages, books, articles, code, and conversations. The "large" refers both to the size of the training data and to the sheer number of parameters (internal settings) in the model. Modern LLMs have hundreds of billions of parameters.

The key breakthrough was discovering that if you train a language model to predict the next word in a sentence on enough data, something remarkable emerges: the model doesn't just learn vocabulary and grammar. It learns facts about the world, how to reason through problems, how to write in different styles, and how to follow instructions.

The Transformer Architecture

All modern LLMs — including GPT-4, Claude, and Gemini — are built on an architecture called the Transformer, introduced by Google researchers in 2017. The key innovation is a mechanism called attention, which lets the model weigh how relevant every word in the input is to every other word when forming a response.

When you ask ChatGPT "What was Einstein's most famous equation and what does it mean?" — the attention mechanism helps the model understand that "it" refers to "equation," that "Einstein" is the person in question, and that "mean" requires an explanation, not just a statement. This is what makes LLMs feel like they genuinely understand language rather than just pattern-matching keywords.

ChatGPT: What It Is and How to Access It

ChatGPT is an AI assistant built by OpenAI, released in November 2022. It became the fastest-growing consumer application in history, reaching 100 million users in just two months. The reason it spread so fast was simple: it was the first time most people had a direct conversation with an AI that felt genuinely intelligent.

How to Get Started (Free)

  1. Go to chat.openai.com
  2. Click "Sign up" and create an account with your email, or sign in with Google.
  3. Verify your email address.
  4. Start your first conversation — it's immediately available.

The free tier gives you access to GPT-4o mini, which is capable for most everyday tasks. ChatGPT Plus (approximately $20/month) gives you access to GPT-4o — significantly more capable for complex reasoning, image analysis, and longer tasks. For most beginners, the free version is more than enough to build real skills.

Understanding GPT Versions

GPT-3.5 was the original model that launched ChatGPT. GPT-4 was a major leap in capability, especially for reasoning and complex tasks. GPT-4o (the "o" stands for "omni") is the current flagship — it's faster, cheaper, and handles text, images, and audio. OpenAI regularly updates these models, so capabilities improve over time.

What ChatGPT Can Do Well — With Real Examples

ChatGPT is a general-purpose AI, which means it's genuinely useful across an unusually wide range of tasks. Here are the most valuable use cases, with concrete examples:

Writing and Editing

Draft emails, blog posts, reports, cover letters, and social media content. Give it a rough outline and it produces a polished draft in seconds. Ask it to "rewrite this more professionally" or "make this shorter and punchier" and it revises accordingly.

Research and Explanation

Ask it to explain complex topics in simple terms. "Explain how mRNA vaccines work to someone with no biology background" will get you a clear, accurate explanation. Useful for research starting points — though you should verify specific facts from primary sources.

Brainstorming and Ideation

Stuck on a project? "Give me 10 creative ways to market a local bakery on a small budget" generates a useful list in seconds. ChatGPT excels at divergent thinking — generating many ideas quickly.

Coding Assistance

Write code from descriptions ("Write a Python function that takes a list of numbers and returns only the even ones"), explain existing code, debug errors, and convert between programming languages. Even non-programmers can use this to automate simple tasks.

Data Analysis and Summarization

Paste a long document and ask for a summary. Extract key points from a lengthy report. Compare two articles. Identify patterns in a list of items. ChatGPT handles large amounts of text far faster than any human can.

What ChatGPT Cannot Do — Critical Limitations

Understanding the limitations is just as important as knowing the capabilities. These are the most important ones:

Knowledge Cutoff

ChatGPT's training data has a cutoff date. It doesn't know about events that happened after that date unless it has been given web browsing tools (available in GPT-4 with the browsing feature). Always confirm time-sensitive information from current sources.

Hallucination

This is the most important limitation. ChatGPT can generate confident, fluent, entirely incorrect information — fabricated statistics, non-existent papers, wrong dates, misattributed quotes. This is called hallucination. It doesn't know when it doesn't know something. Always verify important facts from authoritative sources before using them.

No Memory Between Conversations

By default, each conversation starts fresh. ChatGPT doesn't remember what you discussed yesterday. Within a single conversation, it tracks context — but start a new chat and it's a blank slate. (ChatGPT Plus has a "Memory" feature that can be enabled to address this.)

No Real-World Actions (By Default)

ChatGPT can't send emails, make purchases, or interact with other systems unless specifically configured with tools to do so. It responds with text; it doesn't take actions in the world.

Five Habits That Make You a Power ChatGPT User

Most people who are disappointed with ChatGPT results are simply not using it effectively. These habits will dramatically improve your results:

1. Be Specific and Provide Context

"Write a marketing email" gives ChatGPT almost nothing to work with. "Write a 150-word marketing email to existing customers of a yoga studio announcing a new online membership at $29/month — conversational, warm tone" will produce something immediately useful.

2. Assign a Role

Start your prompt with "Act as a [expert role]." "Act as a financial advisor" or "Act as a senior copywriter with experience in SaaS marketing" shifts the tone, depth, and vocabulary of the response significantly.

3. Iterate, Don't Restart

The first response is a draft, not a final product. Keep the conversation going: "That's good, but make it more concise" or "The second paragraph needs a stronger example." ChatGPT improves dramatically with feedback within the same conversation.

4. Ask for Multiple Options

"Give me 5 different versions of this headline" or "Write three different opening paragraphs" gives you creative options to choose from rather than accepting or rejecting a single output.

5. Verify Critical Information

If you plan to use a specific statistic, quote, date, or claim in professional work — check it independently. Treat ChatGPT outputs as a starting point, not an authoritative source.

Key Takeaways from This Lesson

Large Language Models are deep neural networks trained on billions of text examples to predict and generate language.
ChatGPT is built on OpenAI's GPT architecture; the free tier (GPT-4o mini) is capable for most everyday tasks.
ChatGPT excels at writing, editing, explaining, brainstorming, coding help, and text summarization.
Critical limitations: it can hallucinate facts confidently, has a training knowledge cutoff, and has no memory between chats.
Be specific, assign a role, iterate with feedback, and always verify important facts — these four habits transform your results.

Frequently Asked Questions

ChatGPT is an AI assistant built by OpenAI, powered by a Large Language Model (LLM) called GPT-4. It works by predicting the most likely next token (word fragment) in a sequence based on patterns learned from training on hundreds of billions of words. This produces fluent, contextually relevant responses that often appear intelligent, though the AI doesn't truly "understand" the way humans do.
Yes, ChatGPT has a free tier that gives access to GPT-4o mini. The paid plan (ChatGPT Plus, approximately $20/month) gives access to GPT-4o — the most capable model — along with features like image analysis, web browsing, and higher usage limits. For most beginners, the free tier is an excellent starting point.
ChatGPT is the product — the chat interface you interact with. GPT-4 and GPT-4o are the underlying AI models. ChatGPT can run on different models: the free version uses GPT-4o mini, while ChatGPT Plus uses GPT-4o. GPT-4o is significantly more capable for complex reasoning, longer contexts, and multimodal tasks (text + images).
ChatGPT generates responses based on statistical patterns in its training data, not by looking up verified facts. When it doesn't have reliable information about a topic, it sometimes generates plausible-sounding but incorrect content — this is called hallucination. Always verify important facts, statistics, or claims from authoritative sources before using them professionally.
The most impactful changes are: (1) be specific and detailed in your prompt, (2) provide context about who you are and what the output is for, (3) assign a role ("Act as a..."), (4) iterate — give feedback and ask for revisions rather than accepting the first output, and (5) break complex tasks into smaller steps rather than asking for everything at once.