Reformulates preference learning so the reward model can be folded into the policy in closed form. The result is a single supervised objective over preference pairs, with no separate reward model and no reinforcement learning loop.
@inproceedings{rafailov2023direct,
author = {Rafael Rafailov and Archit Sharma and Eric Mitchell and Stefano Ermon and Christopher D. Manning and Chelsea Finn},
title = {Direct Preference Optimization: Your Language Model is Secretly a Reward Model},
booktitle = {Advances in Neural Information Processing Systems 36 (NeurIPS 2023)},
year = {2023},
eprint = {2305.18290},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2305.18290}
}Ejentic did not author this paper. Credit belongs to the authors named above; the note is ours.
2026
This is the evaluation we run before we let an agent touch anything a customer can see. The finding that matters is the split — a model can pass direct prompt-injection tests and still fold the moment the same instruction comes back through a tool result. Those are different failure surfaces and one test does not cover both.
2026
This is the exact shape of a deployed support or negotiation agent: it works for the business and talks to someone who does not. The split the benchmark finds is the one that matters commercially, because the failure is not leaking data to the counterparty, it is refusing the principal so broadly that the agent stops doing the job it was hired for.
2026
Directly relevant to anyone bolting retrieval onto a model and assuming the citations are the safety mechanism. Adding a fake citation to a true claim made models worse, not better. Grounding that is not verified is worse than no grounding at all, because it arrives with the confidence of evidence.