Understanding AI tokens

Home » Support » Understanding AI tokens

Understanding AI Tokens: A Guide to Token-Based AI Interactions on OnVerb

In the context of AI platforms like OnVerb, tokens play a key role in how users interact with AI models and how usage is measured. This guide will walk you through the concept of tokens, how they are counted, and their significance in managing your interactions with AI services.

What Are Tokens?

Token Definition

In AI model interactions, particularly in natural language processing, a token typically represents a piece of text. This could be a word, a part of a word, or punctuation—essentially any string of characters separated by spaces or other delimiters.

Tokens as Currency

On platforms like OnVerb, tokens are like currency used to measure the amount of textual data the user inputs into the AI system and the output they receive. Each token has a value associated with it, and as you use the AI service, you consume tokens.

Token Count and AI Usage

How Token Count Works

The token count constitutes both the input provided by the user and the output generated by the AI. Here’s how it’s counted:

– Input: The number of tokens you send to the AI, including your prompt and any follow-up questions or interactions.
– Output: The number of tokens in the AI’s response, which can vary based on the complexity and length of the content generated.

Token Economy

Each plan on OnVerb allocates a certain number of tokens to users, essentially setting up an “economy” of how many interactions (inputs and responses) users can have with the AI.

Token Consumption and AI Interactions

Interaction Length and Token Use

Longer interactions or more complex requests will naturally consume more tokens. For instance, asking for a detailed analysis will require more tokens than a straightforward query.

Efficient Token Use

To manage token consumption effectively:

Be Concise: The more precise your input, the less likely you’ll need a lengthy follow-up, saving tokens.
Batch Your Requests: Combine related questions or provide comprehensive prompts to reduce back-and-forth exchanges.
Plan Ahead: Strategize your interactions to get the most value out of each token spent.

OnVerb’s Token System

Token Consumption on OnVerb

OnVerb uses tokens to quantify user interactions with AI models such as GPT 3.5 Turbo and GPT 4 Turbo. Here’s how it translates on OnVerb:

– For GPT 3.5 Turbo:1 chat credit equals 1,000 tokens. (Free on paid subscriptions)
– For GPT 4 Turbo: 4 chat credits are required for every 1,000 tokens.

Managing Your OnVerb Tokens

Track Usage: Monitor your token consumption through the OnVerb platform to plan your activities.
Purchase Add-ons: If needed, you can buy additional tokens in varying amounts to continue engaging with the AI services.