Automated Neuron Labelling Enables Generative Steering and Interpretability in Protein Language Models Paper • 2507.06458 • Published Jul 8