FOUNDATIONS • 27 / 397
Start before LLMs. Understand where AI, ML and neural networks fit.

Loss

Loss is a numerical measure of how wrong a model’s prediction is during training.

Think of it like

Think of Loss as one layer in a very large toolbox: you only need to understand what job that layer performs.

Real life

A phone spam filter, recommendation engine, fraud detector or voice assistant may use Loss somewhere inside.

SRE lens

Training tries to reduce loss across many examples.

Remember thisLoss is a numerical measure of how wrong a model’s prediction is during training.
AIForSREJump to a concept
Search is optional. The main journey is simply ↓ Next.