Data and code for the paper: Finding Safety Neurons in Large Language Models - View it on GitHub