Erase to Improve: Erasable Reinforcement Learning for Search-Augmented LLMs

Open in new window