With the introduction of LLMs, we realised we had to rethink interpretability.