Tutor:
A model doesn’t read letters or words. Before it sees your prompt, a tokenizer chops the text into pieces called tokens, and each piece becomes a number.
Type anything below and watch where the cuts land. Spaces usually stick to the start of the next word.