Context Rot in Large Language Models

Why it matters: Context rot makes LLMs fail long before the window fills. See why, how to measure it, and the context engineering fixes that keep answers reliable.