Yuvion LLM: An Adversarially-Aware Large Language Model for Content And AI Safety
Quick Answer
Yuvion LLM is a new large language model designed for adversarial robustness in AI safety, outperforming larger models like GPT-5.4 on safety benchmarks.
Quick Take
Yuvion LLM is a new large language model designed for adversarial robustness in AI safety, outperforming larger models like GPT-5.4 on safety benchmarks. It employs advanced techniques such as adversarially aware data construction and multi-task safety post-training, demonstrating significant improvements in real-world capability and safety-focused evaluations.
Key Points
- Yuvion LLM integrates adversarially aware data construction and multi-task safety post-training.
- The model includes 93 benchmarks focusing on safety and adversarial robustness.
- Yuvion-8B outperforms many state-of-the-art models on safety tasks.
- Robustness under adversarial conditions is a key feature of Yuvion LLM.
- The model addresses gaps in existing general-purpose model safety approaches.
Paper Resources
📖 Reader Mode
~5 min read
Computer Science > Computation and Language
arXiv:2606.27632 (cs)
Authors:Ting Ma, Xiufeng Huang, Benlei Cui, Xiaowen Xu, Shikai Qiu, Ruijie Jian, Hongxing Li, Guanghui Wang, Longtao Huang, Haiwen Hong, Haolei Xu, Wenjing Jiang, Ziwen Xu, Zhaoyu Fan, Shaoxuan He, Chuxi Xiao, Yujian Li, Xinyue Chen, Chunyang Chai, Wenxuan Liu, Ziheng Wang, Dongjie Zhang, Yangfan Zhou, Libin Dong, Yupeng Cao, Xiaoqian Xia, Jing Wang, Zhe Jiang, Zhenan Ye, Guang Yang, Bin Liu, Wei Peng, Ziqiang Zhu, Meihui Lian, Kaiwen Lv Kacuila, Haidong Ding, Bingyu Zhu, Yan Wang, Hai Zhao, Xuan Jin, Wei Zhao, Pengfei Sun, Wei Wang, Huiming Zhang, Bin Li, Hui Xue
Abstract:As large language models are increasingly deployed in real-world systems, safety failures can still lead to harmful outputs and dangerous misuse. We argue that the essence of safety is adversarial: many failures arise not from natural inputs alone, but from strategic attempts to evade model policies and safeguards. However, existing general-purpose model development largely overlook this adversarial nature, and often remain insufficient for realistic safety scenarios involving planning, tool use, and multi-step reasoning, causing measured safety performance to overestimate real deployment robustness. To address this gap, we present Yuvion LLM, a large language model built for adversarially robust content safety and broader AI safety. Yuvion LLM treats adversarial robustness and agentic capability as first-class objectives. Its pipeline combines adversarially aware data construction, knowledge-enhanced continued pretraining, and policy-grounded multi-task safety post-training, including risk-aware supervised fine-tuning and reinforcement learning-based policy optimization, together with safety-aware agentic reinforcement learning for tool use and multi-step reasoning in complex safety scenarios. We further introduce the Yuvion LLM RiskEval (YLRE), a collection of 93 benchmarks across four evaluation categories, covering diverse open and internal evaluations with a focus on safety, adversarial robustness, and real-world capability requirements. Across these evaluations, Yuvion LLM demonstrates clear advantages on safety-focused benchmarks and particularly strong robustness under adversarial conditions, while maintaining solid overall capability. Notably, Yuvion-8B outperforms most state-of-the-art baselines, including substantially larger models such as GPT-5.4 and Qwen3-MAX, on several safety tasks.
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2606.27632 [cs.CL] |
| (or arXiv:2606.27632v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2606.27632 arXiv-issued DOI via DataCite |
Submission history
From: Benlei Cui [view email]
[v1]
Fri, 26 Jun 2026 01:12:02 UTC (5,130 KB)
Current browse context:
cs.CL
Change to browse by:
Bookmark
Bibliographic and Citation Tools
Code, Data and Media Associated with this Article
Demos
Recommenders and Search Tools
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
— Originally published at arxiv.org
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.CL
See more →Quantifying Prior Dominance in Systems
The study introduces the Normalized Context Utilization (NCU) metric to evaluate Retrieval-Augmented Generation (RAG) systems, revealing that Small Language Models (SLMs) outperform larger models in factual extraction. The findings indicate that traditional scaling laws yield diminishing returns, with a commercial API frequently failing against adversarial evidence due to systemic confidence collapse.
