Parameters are the adjustable values within an AI model that are learned from data during training. In neural networks, these are primarily the weights of connections between neurons. The number of parameters roughly indicates the "size" and capacity of a model – a 70-billion-parameter model can learn more complex patterns than one with 7 billion, but also requires more computing power.
In short
What are parameters in an AI model?
Parameters are the adjustable values of an AI model that are learned from data during training – in neural networks, primarily the connection weights. The number of parameters roughly indicates the size and capacity of a model. More parameters enable more complex patterns but require more computing power.
Using AI in your organisation?
We show you which applications are worth it for your organisation.