Nearby lessons

15 of 30

What is a Token?

Concept in One Line

A token is a small piece of text, such as a word, part of a word, or a symbol, that GPT uses instead of reading full sentences directly.

Why This Concept Is Important

  • GPT does not understand full words like humans.
  • Token count decides cost, speed, and limits.
  • This concept is very important for prompts and interviews.

Key Points to Remember

  • GPT reads tokens, not full sentences.
  • One word can be one token or many tokens.
  • Spaces and symbols also count as tokens.
  • More tokens means more processing.
  • Context window is measured in tokens.

How Tokens Work

Simple Explanation

Example sentence: I am learning AI

Possible tokens:

  • I
  • am
  • learning
  • AI

GPT breaks text into small chunks.

Why Tokens Are Needed

  • Computers understand numbers, not text.
  • Tokens are easy to convert into numbers.
  • Smaller units help the model make better predictions.

Tokens make language machine-friendly.

Simple Flow

Text -> Tokens -> Numbers -> Processing -> Output

Interview / Exam Points

Q1: What is a token in GPT?

A token is a small unit of text used by GPT for processing language.

Q2: Does one word always equal one token?

No. One word can be split into multiple tokens.

Common Confusions to Avoid

  • A token is not always equal to a word.
  • Token count is not the same as character count.
  • Token count affects cost and limits.
  • GPT processes tokens internally.

One Line to Remember

GPT understands text only through tokens.

Quick Self-Check

For Students

GPT reads full sentences directly - Yes / No?

Token count affects context window - Yes / No?

(Correct answers: No, Yes)

🧠 Test Your Knowledge

3 Questions

Progress: 0 / 3
Keep Going!Why Tokenization Is Needed