Researcher, Apple
1 paper at NeurIPS 2025
We examine how problem complexity affects reasoning models' behavior in controlled test settings, revealing key strengths and limitations of their reasoning capabilities.