What People See Is Not AI Itself — Precise Internal Inference, Distorted External Output Through RLHF and Guardrails | LLM News