Researchers in China developed a hallucination correction engine for AI models [cointelegraph]

The “Woodpecker” hallucination correction system can, ostensibly, be applied to any multimodal large language model, according to the research.