Our paper “EasyOPD: An Easy-to-use On-Policy Distillation Framework for Large Language Models” is now available on arXiv (co-authored with NUS and Tencent).