paper-with-me

홈 › Papers

SDPO: Segment-Level Direct Preference Optimization for Social Agents

2025-01-03 · Aobo Kong, Wentao Ma, Shiwan Zhao, Yongbin Li, Yuchuan Wu, Ke Wang, Xiaoqian Liu, Qicheng Li, Yong Qin, Fei Huang

Social agents powered by large language models (LLMs) can simulate human social behaviors but fall short in handling complex goal-oriented social dialogues. Direct Preference Optimization (DPO) has proven effective in aligning LLM behavior with human preferences across a variety of agent tasks. Existing DPO-based approaches for multi-turn interactions are divided into turn-level and session-level methods. The turn-level method is overly fine-grained, focusing exclusively on individual turns, while session-level methods are too coarse-grained, often introducing training noise. To address these limitations, we propose Segment-Level Direct Preference Optimization (SDPO), which focuses on specific key segments within interactions to optimize multi-turn agent behavior while minimizing training noise. Evaluations on the SOTOPIA benchmark demonstrate that SDPO-tuned agents consistently outperform both existing DPO-based methods and proprietary LLMs like GPT-4o, underscoring SDPO's potential to advance the social intelligence of LLM-based agents. We release our code and data at https://github.com/AlibabaResearch/DAMO-ConvAI/tree/main/SDPO.

📄 PDF Abstract BibTeX arXiv:2501.01821

Code (1)

alibabaresearch/damo-convai 공식 구현 pytorch

Similar Papers 제목 키워드 기반

CompassDPO: Dynamics-Controlled Direct Preference Optimization for Robust Safety Alignment

2026-03-07 · Jilong Liu, Yonghui Yang, Pengyang Shao, Wenjian Tao 외 arxiv

Direct Preference Optimization (DPO) has become a standard framework for safety alignment, but its reliance on pairwise preference updates makes training sensitive to imperfect supervision. Existing robust DPO methods of…

Aligning Compound AI Systems via System-level DPO

2025-02-24 · Xiangwen Wang, Yibo Jacky Zhang, Zhoujie Ding, Katherine Tsai 외

Compound AI systems, comprising multiple interacting components such as LLMs, foundation models, and external tools, have demonstrated remarkable improvements compared to single models in various tasks. To ensure their e…

Diffusion-SDPO: Safeguarded Direct Preference Optimization for Diffusion Models

2025-11-05 · Minghao Fu, Guo-Hua Wang, Tianyu Cui, Qing-Guo Chen 외 arxiv

Text-to-image diffusion models deliver high-quality images, yet aligning them with human preferences remains challenging. We revisit diffusion-based Direct Preference Optimization (DPO) for these models and identify a cr…

sDPO: Don't Use Your Data All at Once

2024-03-28 · Dahyun Kim, Yungi Kim, Wonho Song, Hyeonwoo Kim 외

As development of large language models (LLM) progresses, aligning them with human preferences has become increasingly important. We propose stepwise DPO (sDPO), an extension of the recently popularized direct preference…

All

STAR: Mitigating Cascading Errors in Spatial Reasoning via Turn-point Alignment and Segment-level DPO

2026-04-01 · Pukun Zhao, Longxiang Wang, Chen Chen, Peicheng Wang 외 arxiv

Structured spatial navigation is a core benchmark for Large Language Models (LLMs) spatial reasoning. Existing paradigms like Visualization-of-Thought (VoT) are prone to cascading errors in complex topologies. To solve t…

Spatial Reasoning