AI agents are increasingly performing consequential tasks autonomously: writing code, making purchases, and providing advice. But how do we know when to trust them? Current evaluation focuses predominantly on success rates: how often does the agent complete the task? This misses critical questions about how agents behave: Do…
In this talk, Weila Gong will present findings from her book, Implementing a Low-Carbon Future: Climate Leadership in Chinese Cities. Drawing on comparative case studies, process tracing, and over one hundred interviews with policy practitioners inside and outside the government, Gong opens the “black box” of subnational…
This talk focuses on the political economy of venture capital as context for “technological innovation.” Successful startups grow from concepts into companies reaching billions in valuation within a few short years despite making little or no profit. Drawing on ethnographic perspectives from the U.S.
As large language models (LLMs) proliferate, understanding their emergent behaviors remains a central challenge. This talk proposes a framework for interpreting LLM behavior through the lens of geometry. We argue that the structure of an LLM’s internal semantic space—the high-dimensional arrangement of concepts—serves as a…
