STaR introduces a novel two-stage reinforcement learning framework that enables Large Language Models to perform complex cognitive table reasoning with enhan...
Level: advanced
By Unknown
Category: research