1SFT: Instruction-Tuning FoundationsSupervised fine-tuning turns a general language model into a model that follows specific examples of the behavior you want.Advanced 2RLHF & Preference OptimizationHow models learn from human choices, why it works, and where the hard engineering and governance decisions begin.Advanced 3DPO & Simpler AlignmentDirect Preference Optimization turns human preference data into model behavior without the machinery of classic reinforcement learning.Advanced 4Reward Models & Failure ModesA reward model turns human judgment into a training signal, but it also creates new ways for systems to optimize the wrong thing.Advanced 5Alignment Evaluation & SafetyA practical guide to testing whether an AI system behaves as intended, avoids harm, and stays reliable after release.Advanced