Method evidence record

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).

Sources recorded, not reviewed

Source record

Citations copied verbatim from the method’s source record. No claim-level verification is inferred from them.

Direct Preference Optimization: Your Language Model is Secretly a Reward Model

Taxonomic method record · ml-model / deep-learning

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. · URL

Open full method

Curated claims

Claims persisted in the evidence ledger, each with its own assessment.

No curated claims yet

This view does not invent a claim assessment when the ledger has none.

Related methods

Generated from the method graph and shown as machine-suggested relations — no evidence claim is inferred.

Same method familyLatent Diffusion Modelsmachine-suggested · Relational suggestion, not evidence.Same method familyMamba (State Space Model)machine-suggested · Relational suggestion, not evidence.Same method familyMasked Autoencodersmachine-suggested · Relational suggestion, not evidence.Same method familyQLoRAmachine-suggested · Relational suggestion, not evidence.