paper-with-me

Papers

Geometric-Averaged Preference Optimization for Soft Preference Labels

2024-09-10 · Hiroki Furuta, Kuang-Huei Lee, Shixiang Shane Gu, Yutaka Matsuo, Aleksandra Faust, Heiga Zen, Izzeddin Gur

Many algorithms for aligning LLMs with human preferences assume that human preferences are binary and deterministic. However, human preferences can vary across individuals, and therefore should be represented distributionally. In this work, we introduce the distributional soft preference labels and improve Direct Preference Optimization (DPO) with a weighted geometric average of the LLM output likelihood in the loss function. This approach adjusts the scale of learning loss based on the soft labels such that the loss would approach zero when the responses are closer to equally preferred. This simple modification can be easily applied to any DPO-based methods and mitigate over-optimization and objective mismatch, which prior works suffer from. Our experiments simulate the soft preference labels with AI feedback from LLMs and demonstrate that geometric averaging consistently improves performance on standard benchmarks for alignment research. In particular, we observe more preferable responses than binary labels and significant improvements where modestly-confident labels are in the majority.

📄 PDF Abstract BibTeX arXiv:2409.06691

Code (0)

등록된 구현이 없습니다.

Methods 이 논문이 사용한 방법론

DPO 설명 없음

Similar Papers 제목 키워드 기반

Cloud-ScPO: Hidden-State Geometry for Semi-Supervised Preference Optimization in LLM Reasoning

2026-08-02 · Yuzhou Liu, Xiyang Hu arxiv

Preference optimization improves mathematical reasoning in large language models (LLMs), but reliable chosen-rejected pairs usually require verified answers, human annotations, or external reward models. We investigate w…

Mathematical ReasoningPoint Clouds

Soft Preference Optimization: Aligning Language Models to Expert Distributions

2024-04-30 · Arsalan SharifNassab, Saber Salehkaleybar, Sina Ghiassian, Surya Kanoria 외

We propose Soft Preference Optimization (SPO), a method for aligning generative models, such as Large Language Models (LLMs), with human preferences, without the need for a reward model. SPO optimizes model outputs direc…

Computational Efficiency

Learning Where It Matters: Geometric Anchoring for Robust Preference Alignment

2026-02-04 · Youngjae Cho, Jongsuk Kim, Ji-Hoon Kim arxiv

Direct Preference Optimization (DPO) and related methods align large language models from pairwise preferences by regularizing updates against a fixed reference policy. As the policy drifts, a static reference, however, …

On Softmax Direct Preference Optimization for Recommendation

2024-06-13 · Yuxin Chen, Junfei Tan, An Zhang, Zhengyi Yang 외

Recommender systems aim to predict personalized rankings based on user preference data. With the rise of Language Models (LMs), LM-based recommenders have been widely explored due to their extensive world knowledge and p…

Language ModelingLanguage ModellingRecommendation SystemsWorld Knowledge

Disentangling Optimization Scale from Preference Scale in DPO

2026-08-27 · Ivan Kruzhilov arxiv

Direct Preference Optimization (DPO) is a widely used objective for aligning language models from preference data, with the coefficient $β$ commonly interpreted as controlling the KL constraint to a reference policy. We …