TY  - RPRT
TI  - The Best of N Worlds: Aligning Reinforcement Learning with Best-of-N Sampling via max@k Optimisation
AU  - Farid Bagirov
AU  - Mikhail Arkhipov
AU  - Ksenia Sycheva
AU  - Evgeniy Glukhov
AU  - Egor Bogomolov
PY  - 2025
UR  - https://arxiv.org/abs/2510.23393
ID  - 2510.23393
ER  -