LLMs Explained 🧠

What a large language model actually does, in plain terms.

Marcelo Domingues

1 min read

·

Oct 28, 2026

--

Listen

Share

An LLM (Large Language Model) is a neural network trained on huge amounts of text to predict the next token. That’s it. Everything else — answering questions, writing code, summarizing — emerges from doing that one thing extremely well.

How it works

Your text is split into tokens (roughly word-fragments). The model assigns a probability to every possible next token, picks one, appends it, and repeats. A paragraph is just the model running that prediction step hundreds of times.

"The capital of France is" → [Paris: 0.92] [Lyon: 0.01] ... → "Paris"

What it is NOT

Not a database — it doesn’t “look up” facts; it predicts plausible text, so it can be wrong (a “hallucination”). Not deterministic by default — sampling makes outputs vary between runs. Not stateful — each request is independent. You resend the full conversation every time.

Calling one

import anthropic

client = anthropic.Anthropic()
resp = client.messages.create(
    model="claude-opus-4-8",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Explain an LLM in one sentence."}],
)
print(resp.content[0].text)

TL;DR: An LLM is a next-token predictor trained at scale — powerful, but it generates plausible text, not verified truth.

Try one against a real prompt before you trust it for anything that matters. →