Scaling Up RL: Unlocking Diverse Reasoning in LLMs via Prolonged Training

Documenti analoghi

• ProRL: Prolonged Reinforcement Learning Expands Reasoning Boundaries in Large Language Models
  di: Liu, Mingjie, et al.
  Pubblicazione: (2025)
• BroRL: Scaling Reinforcement Learning via Broadened Exploration
  di: Hu, Jian, et al.
  Pubblicazione: (2025)
• ProRL Agent: Rollout-as-a-Service for RL Training of Multi-Turn LLM Agents
  di: Zhang, Hao, et al.
  Pubblicazione: (2026)
• GDPO: Group reward-Decoupled Normalization Policy Optimization for Multi-reward RL Optimization
  di: Liu, Shih-Yang, et al.
  Pubblicazione: (2026)
• HelpSteer2: Open-source dataset for training top-performing reward models
  di: Wang, Zhilin, et al.
  Pubblicazione: (2024)