Records found: 3
'AgentFlow introduces a modular Planner–Executor–Verifier–Generator architecture and Flow-GRPO, a token-level on-policy method that trains only the Planner, reporting substantial gains across ten benchmarks and an open-source MIT implementation.'
ProRLv2 scales RL training to 3,000 steps and combines regularization and exploration techniques to expand reasoning capabilities in compact LLMs, showing strong benchmark gains across math, coding, logic and STEM tasks.
NVIDIA introduces ProRL, a novel reinforcement learning method that extends training duration to unlock new reasoning capabilities in AI models, achieving superior performance across multiple reasoning benchmarks.