Bayes-Adaptive MDPs for Rollout Allocation to Enhance GRPO LLM Training
F. Moradi Kalarde, A. Toroghi, and S. Sanner. "Bayes-Adaptive MDPs for Rollout Allocation to Enhance GRPO LLM Training." Under review, 2026.
Faeze Moradi Kalarde
Postdoctoral Fellow, Department of Computer Science, University of Toronto
- Toronto, Canada
- University of Toronto
- Google Scholar
- GitHub