Exploration and Exploitation Errors Are Measurable for Language Model Agents Rethinking On-Policy Distillation of Large Language Models: Phenomenology, Mechanism, and Recipe Beyond Conservative Automated Driving in Multi-Agent Scenarios via Coupled Model Predictive Control and Deep Reinforcement Learning Evaluating Supervised Machine Learning Models: Principles, Pitfalls, and Metric Selection CodeTracer: Towards Traceable Agent States