IFDECORATOR: Wrapping Instruction Following Reinforcement Learning with Verifiable Rewards Paper • 2508.04632 • Published 2 days ago • 2