Model Architecture

Logits

The raw numerical scores a model produces before they're converted to probabilities.

Definition

Logits are the raw, uncalibrated output of a neural network before any transformation. In the context of language models, the model produces a logit score for every possible next token. These raw scores are then passed through a softmax function to become probabilities. Access to logits is sometimes useful for advanced model analysis or when building custom sampling strategies.

Related Terms

Softmax

The mathematical function that converts a model's raw scores into probabilities.

Token

The basic unit an LLM reads and writes — roughly equivalent to a word or part of a word.

Temperature

A setting that controls how creative or random an AI's outputs are.

Heard enough terminology — ready to talk outcomes?

We translate AI concepts into measurable business results. No upfront fees — you pay only when independently verified results are delivered.

← Back to glossary

Disclaimer

This definition is provided for educational and informational purposes only. It represents a general explanation of a technical concept and does not constitute professional, technical, or investment advice. Artificial intelligence is a rapidly evolving field; terminology, techniques, and capabilities change frequently. Coaley Peak Ltd makes no warranty as to the accuracy, completeness, or currency of the information provided. Nothing on this page should be relied upon as the sole basis for commercial, technical, legal, or investment decisions without independent professional advice.

Document reference: ISO_webpage_knowledge-base_glossary_v1

Last modified: 29 March 2026

Knowledge Base·Model Architecture·Logits