Large reasoning models form an accurate Sierpinski-triangle world model of the Tower of Hanoi at the prompt, then lose it during extended reasoning, and restoring it via activation steering improves planning accuracy.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.AI 1years
2026 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Transformers Struggle to Use Their Emergent World Models: Revisiting the Tower of Hanoi, and the Illusion of Thinking
Large reasoning models form an accurate Sierpinski-triangle world model of the Tower of Hanoi at the prompt, then lose it during extended reasoning, and restoring it via activation steering improves planning accuracy.