Post-training of language models, either through reinforcement learning, preference optimization or supervised finetuning, tends to sharpen the output probability distribution and reduce the diversity of generated responses. This is particularly a problem for creative generative tasks where varied responses are desired. In this work we introduce Diverse Preference Optimization (DivPO), an optimization method which learns to generate much more diverse responses than standard pipelines, while maintaining the quality of the generations. In DivPO, preference pairs are selected by first considering a pool of responses, and a measure of diversity among them, and selecting chosen examples as being more rare but high quality, while rejected examples are more common, but low quality. DivPO results in generating 45.6% more diverse persona attributes, and an 74.6% increase in story diversity, while maintaining similar win rates as standard baselines.
View on arXiv@article{lanchantin2025_2501.18101,
title={ Diverse Preference Optimization },
author={ Jack Lanchantin and Angelica Chen and Shehzaad Dhuliawala and Ping Yu and Jason Weston and Sainbayar Sukhbaatar and Ilia Kulikov },
journal={arXiv preprint arXiv:2501.18101},
year={ 2025 }
}