1 paper
Lucas Torroba Hennigen, Shannon Shen, Aniruddha Nrusimha +3
LLMs are vulnerable to hallucinations, and thus their outputs generally require laborious human verification for high-stakes applications. To this end, we propose symbolically grou…