You can see some examples of the output of the ML model in the scrolling threat feed on our homepage here: https://socket.dev

We use LLMs both for detecting threats as well as explaining the output of traditional static analysis in a way that makes the findings understandable to the average developer.