Parameter
Full Form: Model Parameter
Category: AI Fundamentals
📖 Definition
Parameters are the internal values in an AI model that are learned during training. They're the 'knowledge' the model uses to generate outputs. Larger models have more parameters.
🔑 Key Points
- Billions of parameters in modern LLMs (GPT-4: ~1.7 trillion)
- Parameters are adjusted during training to minimize errors
- More parameters generally means more capable (but not always)
- Also called 'weights' in neural networks
💡 Why It Matters
Parameter count is often used to describe model size. Understanding parameters helps you compare different AI models.