Picture for Xiangru Peng

Xiangru Peng

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

Add code
Jun 26, 2024
Viaarxiv icon