Saddle scrambling lets reasoning models explore more solutions, measured by Lyapunov indicator
wgilpin0 · x · 2026-09-09
The team explains the mechanism: high-dimensional dynamics act like a Plinko game — trajectories start nearby and converge to the same answer, but saddles scatter them along different routes via incorrect answers. 'Saddle scrambling' lets reasoning models check more possible solutions. They compute the fast Lyapunov indicator to measure transient chaos, finding it correlates with how many solutions the model visits before converging.
More from Research
- DeepMind paper: LLMs can't sample randomly on their own, only convert given seeds — rohanpaul_ai · 2026-10-27
- DeepMind paper: LLMs can't sample randomly on their own — pass random numbers from code instead — rohanpaul_ai · 2026-10-27
- World Model Assigns Weight and Friction to Objects, Then Simulates with a Physics Engine — jonstephens85 · 2026-10-27
- Token Collapse: Study Finds Graph Soft Prompting Tokens in LLMs Fail to Encode Structure — Sapienza · 2026-10-27
- Mathathon 2.0 launches with Terence Tao's SAIR, $1.5M in compute for teams — _sathvikr · 2026-10-27
- The 'sweet lesson' is real: two kinds of ML problems need opposite strategies — mike64_t · 2026-10-27