1 paper · 1 filter
Pranjal Garg
Probes are the workhorse of interpretability. If a model's hidden states predict a variable, the model is said to represent it. But a probe score has no fixed meaning. An R2 of…