← The notebook
Method2025 · 07 · 30#interpretability

Activation patching to find the cause

Probes tell you what is represented. Patching tells you what matters. By swapping one activation and watching the effect, you move from 'this information is present' to 'this component causes the behavior'.

Sweep the patch across layers and positions to localize the circuit. The spots where patching flips the answer are the load-bearing ones; everything else is along for the ride.

The closest thing in this field to a controlled experiment.