Zubnet AILearnWiki › Tokenizer
Fundamentals

Tokenizer

Tokenization
The algorithm that converts raw text into tokens before a model can process it. A tokenizer maintains a fixed vocabulary of token types and splits any input text into a sequence of those tokens. Different models use different tokenizers — the same sentence tokenizes differently for Claude, GPT, and Llama, which affects context usage and cost.

Why it matters

The tokenizer is the invisible layer between your text and the model. It determines how many tokens your prompt costs, why some languages are more expensive than others, and why code sometimes uses context faster than prose. When you hit a context limit or see unexpected API costs, the tokenizer is usually the explanation.

Deep Dive

Most modern tokenizers use Byte Pair Encoding (BPE) or a variant called SentencePiece. BPE works by starting with individual bytes or characters and repeatedly merging the most frequent adjacent pair into a new token. After thousands of merges, common words like "the" become single tokens, while rare words get split into subword pieces. The word "tokenization" might become ["token", "ization"] or ["token", "iz", "ation"] depending on the specific merge table.

Vocabulary Size Matters

A tokenizer's vocabulary size is a real engineering trade-off. Larger vocabularies (100K+ tokens) compress text more efficiently — common words and phrases get dedicated tokens, so less context is consumed. But larger vocabularies also mean a bigger embedding table at the model's input and output layers. For a model with dimension 4096, each vocabulary entry adds 4096 parameters to both the embedding and the unembedding layers. At 128K vocabulary, that's over a billion parameters just for the token tables. Smaller models feel this overhead proportionally more.

The Multilingual Tax

Tokenizers are trained on a corpus, and the language distribution of that corpus determines efficiency. English text typically tokenizes at roughly 1 token per word. But languages like Chinese, Japanese, Korean, Arabic, and Hindi can require 2–4x more tokens for equivalent meaning, because their characters appear less frequently in English-dominated training data and earn fewer dedicated merges. This isn't just an academic concern — it means non-English users pay more per API call and fit less content in the context window. Some newer tokenizers (like Llama 3's) explicitly train on more balanced multilingual data to reduce this gap.

Tokenizer Artifacts

Quirks in tokenization explain several LLM behaviors people find puzzling. Models struggle with character-level tasks (counting letters in "strawberry") because they see tokens, not characters. They handle some variable names better than others because common names like "result" are single tokens while unusual ones fragment. They sometimes produce slightly different outputs for semantically identical inputs because the token boundaries differ. Understanding the tokenizer helps you understand the model.

Related Concepts

← All Terms
← Token Tool Use →