Fetching the paper…

Step-DPO: Step-wise Preference Optimization for Long-chain Reasoning of LLMs · Around