Cengiz Pehlevan: Solvable Models of Test-Time Scaling
Abstract
Modern generative models increasingly spend compute at inference time. When does this extra compute improve performance, when does it saturate, and when can it backfire? This talk introduces a family of exactly solvable high-dimensional models, analyzed using tools from random matrix theory, that make these questions analytically tractable and offer a quantitative theory of test-time scaling.
Video
Chat is not available.
Successful Page Load