Machine learningDeep Learning, Language Models, RLHF Alternatives
Direct Preference Optimization
Direct Preference Optimization (DPO) is a training method introduced by Rafailov et al. in 2023 that aligns language models with human preferences without requiring an explicit reward model. By directly optimizing for preference pairs (better response vs worse response), DPO simplifies the training pipeline compared to reinforcement learning from human feedback (RLHF).
Open in MethodMindSoonVideoSoon
Read the full method
Members only
Sign inSign in with a free account to read this section.
Sources
- Rafailov, R., Sharma, A., Mitchell, E., Manning, C. D., Ermon, S., & Finn, C. (2023). Direct preference optimization: Your language model is secretly a reward model. arXiv preprint arXiv:2305.18290. link ↗