Expanded Explanation
Large Language Models (LLMs) are neural networks — usually transformers — trained on billions of tokens of text. They can answer questions, summarise, translate, reason and generate code. Examples include GPT-4, Claude and Gemini.
Technical explanation
LLMs are autoregressive transformer models trained with next-token prediction, then aligned via supervised fine-tuning and reinforcement learning from human feedback (RLHF).
Simple Examples
- GPT-4 in ChatGPT
- Claude 3
- Gemini 1.5