Cal-DPO: Calibrated Direct Preference Optimization for Language Model Alignment

Devs

Cal-DPO: Calibrated Direct Preference Optimization for Language Model Alignment | Read Paper on Bytez