Fetching the paper…

Enhancing Multi-Step Reasoning Abilities of Language Models through Direct Q-Function Optimization · Around