Researchers at MIT developed a new algorithm called Co-LLM. Pairing of LLMs through Co-LLM led to more accurate results.