DefensesarXiv 2609.39820
FailBank：用运行时反馈自进化提升 VLA 模型任务成功率
提出 FAIL BANK，用运行时 CBF 反馈自进化更新 VLA 策略
- arXiv
- 2609.39820
- Published
- Evaluated models
- π0.5, π0
- DefensesModel hardening
- RiskDangerous agent actions
All papers
Filter2
Model risks
Impact
System type
Select any value within a category; different categories must all match. Counts include the other filters you have selected.Another 265papers have no tags yet and are not included in filtered results.
Another 265papers do not have category tags yet and are not included in filtered results.
AgentBound：面向工具型 LLM Agent 安全的四路反事实评测框架
提出 AGENT BOUNDARY，用四向反事实任务评测工具智能体行动边界
- arXiv
- 2609.33658
- Published
- Layer
- Application layer
- Setting
- AI Agent
- Evaluated models
- GPT-5.5, Claude Opus 4.8, Gemini 3.5 Flash and others15
Abstract四向反事实基准把过度拒绝和未授权遵从拆开，Thought 校正在十个配置上提高已授权完成。
AGENT BOUNDARY 是面向工具使用 LLM 智能体的四向可执行反事实评测，并附带一个决策时 Thought 校准模块。
研究用 30 项漏洞分析任务评测自主安全智能体的安全对齐效果
评测自主安全智能体在授权漏洞分析中的对齐拒答效应
- arXiv
- 2605.19722
- Published
- Layer
- Application layer
- Setting
- coding agent
- Evaluated models
- Gemma 4 31B official instruction-tuned model, TrevorJS Gemma 4 31B uncensored GGUF, Gemma 4 26B A4B official instruction-tuned model and others8
Abstract轨迹级评测把拒答、接地、工具接口分开。
这项工作把安全对齐效应测成自主安全智能体的轨迹属性，而不是单轮拒答率。
- DefensesarXiv 2601.12672
VILTA: Incorporating VLM into closed-loop training to improve driving policy robustness
提出 VILTA，在闭环训练中用 VLM 编辑邻车轨迹以提升驾驶策略鲁棒性
- arXiv
- 2601.12672
- Published
- DefensesModel hardening
- RiskDangerous agent actions
AbstractVILTA 在闭环中编辑邻车轨迹以训练驾驶策略。
VILTA 是一个把视觉语言模型放进自动驾驶闭环训练的框架，用来应对长尾危险场景不足。
You have reached the end