Inference-time alignment provides an efficient alternative for aligning LLMs with humans. However, these approaches still face challenges, such as limited scalability due to policy-specific value functions and latency during the inference phase. In this paper, we propose a novel approach, Diffusion-styled Preference Optimization (\model), which provides an efficient and policy-agnostic solution for aligning LLMs with humans. By directly performing alignment at sentence level, \model~avoids the time latency associated with token-level generation. Designed as a plug-and-play module, \model~can be seamlessly integrated with various base models to enhance their alignment. Extensive experiments on AlpacaEval 2, MT-bench, and HH-RLHF demonstrate that \model~achieves superior alignment performance across various settings, achieving a favorable trade-off between alignment quality and inference-time latency. Furthermore, \model~demonstrates model-agnostic scalability, significantly improving the performance of large models such as Llama-3-70B.
View on arXiv@article{chen2025_2503.04240,
title={ DiffPO: Diffusion-styled Preference Optimization for Efficient Inference-Time Alignment of Large Language Models },
author={ Ruizhe Chen and Wenhao Chai and Zhifei Yang and Xiaotian Zhang and Joey Tianyi Zhou and Tony Quek and Soujanya Poria and Zuozhu Liu },
journal={arXiv preprint arXiv:2503.04240},
year={ 2025 }
}