Showing cs.LGShow all
2 papers · 1 filter
cs.LG2026
Spend Your Rollouts Where It Counts: Rollout Allocation for Group-Based RL Post-Training
Woojeong Kim, Ziyi Yang, Jing Nathan Yan +1
Reinforcement learning (RL) is the dominant paradigm for post-training large language models. However, in the online, on-policy setting, rollout generation dominates the computatio…
cs.LG2025
Mutual-Taught for Co-adapting Policy and Reward Models
Tianyuan Shi, Canbin Huang, Fanqi Wan +5
During the preference optimization of large language models (LLMs), distribution shifts may arise between newly generated model samples and the data used to train the reward model…