# What is Temperature? Definition & Technical Mechanics

> **Source:** [SPIDITS AI AI Glossary](https://spidits.com/ai-glossary)  
> **Category:** Model Operations  
> **Canonical Citation:** [https://spidits.com/ai-glossary/temperature](https://spidits.com/ai-glossary/temperature)

## Definition
Temperature is a parameter that controls the randomness and creativity of text generated by an autoregressive language model during inference. Higher values increase randomness, while lower values make outputs more deterministic.

## Architectural Mechanics & Deep Dive
Temperature is a configuration parameter that controls the randomness and creativity of generative model outputs. By scaling the logits before the softmax calculation, a low temperature (near 0) forces deterministic, high-probability token selections, while a high temperature (near 1) increases token diversity.

## Industry Applications & Real-World Use Cases
Response tuning, generation formatting control, and brainstorming triggers

## Frequently Asked Questions

### What happens at Temperature = 0?
The model becomes deterministic, choosing the single most probable word token at every step (greedy decoding).

### When should you use a high temperature?
During creative writing tasks, brainstorming sessions, or scenario generation where diverse outputs are desired.

## Related Terminology
- [llm](https://spidits.com/ai-glossary/llm)
- [inference](https://spidits.com/ai-glossary/inference)
- [tokenization](https://spidits.com/ai-glossary/tokenization)

---
*From the SPIDITS AI Knowledge Base (336+ terms indexed). Explore full technical definitions: [https://spidits.com/ai-glossary](https://spidits.com/ai-glossary)*