Automated Neuron Labelling Enables Generative Steering and Interpretability in Protein Language Models

Open in new window