Interpretability Happy Hour w. Goodfire
Theme for this month’s happy hour is ML engineering and infra for interpretability. See our live-steering demo on a trillion-parameter coding agent: we'll inject semantic vectors into Kimi K2 Thinking in real time (mid-chain-of-thought) to steer for things like concision or Gen Z thinking style. A follow-up demo will show one of our research agents in action!
Come hang out with the Goodfire team, talk to our engineers about our custom inference engine and core library that enables it, and learn about our open ML engineering roles!
Apps & drinks provided!
Goodfire is an AI interpretability research company focused on understanding and intentionally designing the next frontier of safe and powerful AI systems. Find all our open engineering positions on the Careers page.