# What is GPT? Definition & Technical Mechanics

> **Source:** [SPIDITS AI AI Glossary](https://spidits.com/ai-glossary)  
> **Category:** Neural Architectures  
> **Canonical Citation:** [https://spidits.com/ai-glossary/gpt](https://spidits.com/ai-glossary/gpt)

## Definition
GPT (Generative Pre-trained Transformer) is a decoder-only autoregressive transformer architecture developed by OpenAI. It was pre-trained on massive text datasets to predict next words, pioneering the modern conversational AI era.

## Architectural Mechanics & Deep Dive
GPT (Generative Pre-trained Transformer) is a family of autoregressive decoder-only Transformer models pioneered by OpenAI. Trained to predict the next token on massive text datasets, GPT models serve as the foundation for modern conversational assistants, establishing the power of scaling parameters, compute, and data to achieve advanced reasoning capabilities.

## Industry Applications & Real-World Use Cases
Interactive text generation, code editing, and contextual reasoning API integrations

## Frequently Asked Questions

### What does "Pre-trained" mean in GPT?
It means the model has already been trained on massive datasets to understand basic grammar, facts, and logic before being tuned for specific tasks.

### Why is GPT a decoder-only architecture?
Because its primary purpose is text generation. It uses causal attention masking to prevent tokens from looking at future tokens in a sequence.

## Related Terminology
- [llm](https://spidits.com/ai-glossary/llm)
- [transformer](https://spidits.com/ai-glossary/transformer)
- [autoregressive-model](https://spidits.com/ai-glossary/autoregressive-model)

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