GPT-4: Researchers have developed a method called LLM+P that enhances large language models (LLMs) by connecting them to general-purpose planners and calculators. This allows LLMs to accurately solve planning problems and answer questions without changing the LLMs themselves. The LLM+P method can be applied to any class of problems with a comprehensive solver, improving the functional competence of LLMs like GPT-4 and ChatGPT.
Read more at MarkTechPost…