From Hallucination to Certainty: Meta-Knowledge Guided Self-Correcting Large Language Models | Synapse