Preprint
Aug 2026
Forking Fast: Efficiently Estimating Uncertainty Dynamics in Text Generation
It is shown that when resampling many reasoning chains, uncertainty dynamics converge to stable patterns, and noise is largely an artifact of sampling rather than an LLM's sensitivity to each individual token or reasoning step.
Eric J. Bigelow, Amir Zur, Satchel Grant et al.
· 1 citation