Layer-of-Thoughts (LoT) prompting filters LLM outputs through hierarchical layers, like a smart document sieve.