Ryan Park
Ryan Park has 2 papers in our library, in 2024, most of them on Language models & pretraining; often with Chelsea Finn, Rafael Rafailov, Stefano Ermon.
No OpenAlex profile matched yet.
📄 2 works in our library · 🔗 Cited 155 times in our library · ⭐ Paper Stars ranking: #6728
Fields, topics and concepts
📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Artificial Intelligence · 2 worksArts and Humanities · 2 worksComputer Science · 2 worksLanguage and Linguistics · 2 works
🧭 Topics 1Research subjects and languages the work is about.Language models & pretraining · 1 work
Works by year
2020s
Papers
- Disentangling Length from Quality in Direct Preference Optimization (2024 · 105 citations)
The paper looks at verbosity exploitation in Direct Preference Optimization (DPO). Classical RLHF is known to push models toward longer answers, but this had not been studied for DPO, which trains offline and has no explicit reward model. Existing length-control methods therefore cannot be applied to it directly. - From r to Q*: Your Language Model is Secretly a Q-Function (2024 · 72 citations)
Direct Preference Optimization (DPO) is normally derived as a contextual bandit, where a whole response counts as one action.
Co-authors here
Venues
- Findings of the Association for Computational Linguistics: Findings of the Association for Computational Linguistics 2024
Cited by
Updated
This page in Markdown → https://md.research.elamur.ai/free/en/authors/ryan-park/