In AI, “alignment” is the thin layer that nudges a raw model toward a chosen set of values - think guard-rails, etiquette, mission statement, all baked into its training loop. So Alignment is kind of a religion for LLM?