---
schema: 1
type: term
title: '**token**'
description: >-
  **token** — the basic unit of text (a word, part of a word, or punctuation
  mark) that a language model processes and produces.
canonical: 'https://liquidgold.cc/channel/everyinc/explore/term/token-c84c147a59ee'
markdown: 'https://liquidgold.cc/channel/everyinc/explore/term/token-c84c147a59ee.md'
llms: null
parent: 'https://liquidgold.cc/channel/everyinc/llms.txt'
channel: everyinc
updated: '2026-10-23T22:12:35.346Z'
language: null
children: []
---


# \*\*token\*\*

Canonical: https://liquidgold.cc/channel/everyinc/explore/term/token-c84c147a59ee

Parent: [Channel index](https://liquidgold.cc/channel/everyinc/llms.txt)

## Definition

**token** — the basic unit of text (a word, part of a word, or punctuation mark) that a language model processes and produces.

In the captured material, which explores a large language model (Anthropic's unreleased Claude Mythos) and how its capabilities are discussed [^4b2e7916-64b0-5a56-95fc-903f3587ba28@0], a **token** is the basic unit of text that a language model reads and generates — a word, part of a word, or piece of punctuation. The material discusses the model in terms of *capability* and *release decisions* rather than its internals, so **token** appears here in its standard language-model sense rather than as a concept the video itself defines or elaborates on.