Can Large Language Models Act as Symbolic Reasoners?

Rob Sullivan, Nelly Elsayed

The performance of Large language models (LLMs) across a broad range of domains has been impressive but have been critiqued as not being able to reason about their process and conc…