Adaptive Tool Generation with Models as Tools and Reinforcement Learning

Chenpeng Wang, Xiaojie Cheng, Chunye Wang +2

Tool-augmented language models have demonstrated strong capabilities, but their reliance on live API access creates scalability and reliability challenges during training and deplo…