arXiv cs.CL· Jiawei Chen, Ruoxi Xu, Boxi Cao, Zhuoqun Li, Ruotong Pan, Yunfei Zhang, Zhirui Yang, Jingyan Chen, Tingting Gao, Han Li, Yaojie Lu, Xianpei Han, Le Sun, Xiangyu Wu, Hongyu Lin·· 13 小时前AI 评分31
OmniSimulator：对齐小语言模型实现真实异构行为建模
OmniSimulator: Aligning Small Language Models for Authentic Heterogeneous Behavior Modeling
AI 导读
OmniSimulator 训练框架通过五维结构化 schema 组织行为导向推理，并用教师模型生成高质量推理轨迹，提升小模型在短视频平台用户模拟中的预测能力。基于 200 名真实用户三个月交互轨迹构建 9,881 条训练实例，使 Qwen3-8B 平均得分提升 34.33%，性能可与 Claude-Sonnet-4.5 竞争，并支持更小规模的自托管部署。
正文 · 原文
Authors:Jiawei Chen, Ruoxi Xu, Boxi Cao, Zhuoqun Li, Ruotong Pan, Yunfei Zhang, Zhirui Yang, Jingyan Chen, Tingting Gao, Han Li, Yaojie Lu, Xianpei Han, Le Sun, Xiangyu Wu, Hongyu Lin
Abstract:Large language models (LLMs) have shown great potential as general-purpose user simulators for interactive systems. Prior studies have demonstrated that training on high-quality data can substantially improve model reasoning and downstream task performance. However, the upper bound of user simulation is still constrained by the model's reasoning paradigm: native chain-of-thought often degenerates into shallow pattern matching and fails to adequately model the behavioral factors underlying real user decisions. In this paper, we propose OmniSimulator, a novel training framework designed to improve the predictive ability of small LLMs for realistic user simulation on short-video platform environments. Specifically, OmniSimulator organizes behavior-oriented reasoning through a structured schema with five dimensions, carefully selects data from real user-behavior scenarios, and introduces a teacher model to generate high-quality reasoning traces under the proposed schema, helping small models learn behavior-relevant reasoning patterns rather than relying on superficial behavioral correlations. Based on this process, we construct 9,881 high-quality training instances from three months of historical interaction trajectories of 200 real users, averaging about 50 distilled examples per user. Experimental results show that OmniSimulator improves the average score of Qwen3-8B by 34.33%, achieves performance competitive with Claude-Sonnet-4.5, and supports smaller-scale, self-hosted deployment. These results suggest that learning structured behavior-oriented reasoning patterns is a promising direction towards scalable and high-fidelity LLM-based user simulation.
| Comments: | Accepted by NeurIPS 2026 |
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.29472 [cs.CL] |
| (or arXiv:2610.29472v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.29472 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Jiawei Chen [view email]
[v1]
Wed, 21 Oct 2026 09:30:12 UTC (548 KB)
来源：arXiv cs.CL · arxiv.org