TY  - RPRT
TI  - ReProbe: Efficient Test-Time Scaling of Multi-Step Reasoning by Probing Internal States of Large Language Models
AU  - Jingwei Ni
AU  - Ekaterina Fadeeva
AU  - Tianyi Wu
AU  - Mubashara Akhtar
AU  - Jiaheng Zhang
AU  - Elliott Ash
AU  - Markus Leippold
AU  - Timothy Baldwin
AU  - See-Kiong Ng
AU  - Artem Shelmanov
AU  - Mrinmaya Sachan
PY  - 2026
UR  - https://arxiv.org/abs/2511.06209
ID  - 2511.06209
ER  -