Introduced on day 40
Linear probe
A simple linear or logistic model trained on frozen pretrained features; both a cheap baseline and a first step before fine-tuning.
In detail
It measures how useful the pretrained features are for your task without changing them. Starting a fine-tune from a trained probe, rather than a random head, avoids early large updates that distort the features.