Yuvion LLM: An Adversarially-Aware Large Language Model for Content And AI Safety
作者: 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
分类: cs.CL
发布日期: 2026-06-26
💡 一句话要点
提出Yuvion LLM以解决内容与AI安全问题
🎯 匹配领域: 支柱二：RL算法与架构 (RL & Architecture) 支柱九：具身大模型 (Embodied Foundation Models)
关键词: 大型语言模型 对抗性鲁棒性 AI安全 多任务训练 强化学习
📋 核心要点
- 现有大型语言模型在安全性方面存在不足，尤其是在对抗性输入和复杂场景下的表现不佳。
- Yuvion LLM通过对抗性鲁棒性和代理能力的结合，采用多种训练策略提升模型的安全性和实用性。
- 实验结果表明，Yuvion LLM在安全任务上优于大多数现有基线，包括更大规模的模型，如GPT-5.4和Qwen3-MAX。
📝 摘要（中文）
随着大型语言模型在现实系统中的广泛应用，安全失效仍可能导致有害输出和危险误用。本文认为安全的本质是对抗性的：许多失效并非仅源于自然输入，而是来自于规避模型政策和保护措施的战略尝试。现有的通用模型开发往往忽视这种对抗性，导致在涉及规划、工具使用和多步推理的现实安全场景中表现不足。为此，本文提出Yuvion LLM，一个专为对抗性鲁棒内容安全和更广泛的AI安全构建的大型语言模型。Yuvion LLM将对抗鲁棒性和代理能力视为首要目标，结合了对抗性数据构建、知识增强的持续预训练和基于政策的多任务安全后训练等方法。
🔬 方法详解
问题定义：本文旨在解决大型语言模型在现实应用中面临的安全性挑战，尤其是对抗性输入导致的失效问题。现有方法往往忽视对抗性特征，导致安全性能被高估。
核心思路：Yuvion LLM将对抗鲁棒性和代理能力作为首要目标，通过对抗性数据构建和多任务训练策略，提升模型在复杂安全场景下的表现。
技术框架：Yuvion LLM的整体架构包括对抗性数据构建、知识增强的持续预训练、政策驱动的多任务安全后训练，以及风险感知的监督微调和基于强化学习的政策优化。
关键创新：Yuvion LLM的主要创新在于将对抗性鲁棒性与代理能力结合，形成一个综合的安全训练框架，显著提升了模型在对抗性条件下的表现。
关键设计：在模型设计中，采用了风险感知的损失函数和强化学习策略，以确保模型在多步推理和工具使用中的安全性和有效性。具体参数设置和网络结构细节在论文中进行了详细描述。
🖼️ 关键图片
📊 实验亮点
在多项安全基准测试中，Yuvion LLM表现出明显优势，尤其是在对抗性条件下的鲁棒性。Yuvion-8B在多个安全任务上超越了大多数现有基线，包括更大规模的模型，如GPT-5.4和Qwen3-MAX，显示出其在安全性方面的显著提升。
🎯 应用场景
Yuvion LLM的研究成果在多个领域具有潜在应用价值，包括内容生成、自动化客服、教育辅助等。其对抗性鲁棒性和安全性设计能够有效降低模型在实际应用中的风险，提升用户信任度。未来，该模型的技术可以进一步推广到更广泛的AI系统中，增强整体安全性。
📄 摘要（原文）
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.