About this document

ShojaeeEtAl IllusionOfThinkinginLLMs 2025 by juliakatten05 is a document available to read on EtoBox.

The paper investigates the strengths and limitations of Large Reasoning Models (LRMs) in relation to problem complexity, revealing that while LRMs show improved performance on reasoning tasks, they face significant accuracy collapse beyond certain complexity thresholds. The study employs controllable puzzle environments to analyze both final answers and internal reasoning processes, highlighting inefficiencies in LRMs

Author

juliakatten05

Language

EN