Adversarial Machine Learning

Nicholas Carlini joined Bryan, Adam, and the Oxide Friends to talk about his work with adversarial machine learning. He's found sequences of--seemingly random--tokens that cause LLMs to ignore their restrictions! Also: printf is Turing complete?!
In addition to Bryan Cantrill (mastodon.social/@bcantrill) and Adam Leventhal (mastodon.social/@ahl), we were joined by special guest Nicholas Carlini.
If we got something wrong or missed something, please file a PR! Our next show will likely be on Monday at 5p Pacific Time on our Discord server; stay tuned to our Mastodon feeds for details, or subscribe to this calendar (calendar.google.com/calendar/ical/c_318925f418…). We'd love to have you join us, as we always love to hear from new speakers!