Adaptive $Q$-Aid for Conditional Supervised Learning in Offline Reinforcement Learning | Read Paper on Bytez