Tilde Research @ Stanford

​Join us for a conversation on where model architecture research is headed.

​We’ll share a broader look at some of the directions we’re most excited about in architecture design and how they connect to questions about reasoning, efficiency, and search. We’ll also explore how different ways of structuring models open up new behaviors, failure modes, and opportunities for control, and how this shapes the kinds of research directions we think are most promising.

​Afterwards, we’ll open things up for Q&A, followed by an interactive activity exploring open problems in model architectures and training dynamics.

This event is open to all Stanford students. Pizza and boba will be served after the event.

​Tilde Research is a moonshot AI lab advancing mechanistic interpretability, new architectures, and pretraining science. We’re backed by Khosla Ventures, Yufei Zhao, and researchers at OpenAI and Anthropic. For more information, please visit tilderesearch.com.