Fetching the paper…

INTELLECT-2: A Reasoning Model Trained Through Globally Decentralized Reinforcement Learning · Around