Let it Calm: Exploratory Annealed Decoding for Verifiable Reinforcement Learning

Written on October 6, 2025