Interpretability for Security1Reading the Machine: Features, Sparse Autoencoders, and What Interpretable Actually Buys a DefenderWatch2Probes as Detectors: Classifying Jailbreak, Refusal, and Unsafe States From ActivationsWatch3Detecting Deception From the Inside: Promise, Evidence, and the Adaptive-Adversary ProblemWatch4Runtime Activation Monitoring: Latency, Coverage, and the Cost of Watching InternalsWatch5Interpretability's Limits as a Control: False Negatives, Goodharting the Probe, and Layered DefenseWatch