Translation in progress Some entries still include text in the original language.

Oct 12 · Monday

All papers

Found 10 papers

Filter5

Model risks

Deception and scheming10 Reward hacking8 Misalignment and value divergence4 Dangerous agent actions30 Hallucination and fabrication3

Defense type

Instruction–data separation1 Permissions and approval1 Monitoring and auditing9

Impact

Oversight failure7

System type

AI Agent10 coding agent6

Layer

Model layer5 Application layer10

System component

Monitors9

Select any value within a category; different categories must all match. Counts include the other filters you have selected.Another 233papers have no tags yet and are not included in filtered results.

Model risks：Deception and scheming Defense type：Instruction–data separation Defense type：Permissions and approval Defense type：Monitoring and auditing Layer：Application layer Clear all

Another 233papers do not have category tags yet and are not included in filtered results.

DefensesarXiv 2610.12445Oct 9
FAR.AI paper: white-box probes surpass Opus 5.5 text monitoring on SHADE-Arena with 98.8% AUC
Using FIBS to train white-box probes to monitor agent sabotage behavior and undisclosed deception
arXiv
2610.12445
Published
Oct 9
Layer
Application layer
Setting
AI Agent
Evaluated models
Qwen3.5-397B-A17B, Qwen3.5-122B-A10B, Qwen3.5-27B and others10
- DefensesMonitoring and auditing
- RiskDeception and scheming
- ComponentMonitors
Paper analysisRead analysis
Evaluation and benchmarksarXiv 2610.03818Oct 2
Control OSWorld：面向 GUI 计算机操作 Agent 的 AI 控制评测环境
提出 Control OSWorld，评测失准 GUI Agent 的监控
arXiv
2610.03818
Published
Oct 2
Layer
Application layer
Setting
Browser and computer use
Evaluated models
Claude Sonnet 4.6, Claude Opus 4.8, GPT-5.2 and others6
- DefensesMonitoring and auditing
- RiskDeception and scheming
- ComponentMonitors
Paper analysisRead analysis
DefensesarXiv 2609.36170Sep 29
Assay：用内容寻址证据图约束 AI 编码智能体的声明失效
提出 Assay，用依赖锥哈希绑定代码断言并在合并门禁拦截失效声明
arXiv
2609.36170
Published
Sep 29
Layer
Application layer
Setting
coding agent
- DefensesPermissions and approval
- RiskDeception and scheming
- ComponentPermissions and sandboxing
AbstractAssay 提供随依赖内容失效的证据管理。
Assay 是一套面向 AI 辅助软件交付的内容寻址证据系统，关注代码改变后，测试通过、无秘密或行为保持等断言还能否用于批准合并。
Paper analysisRead analysis
DefensesarXiv 2609.19587Sep 17
红队测试 Auto Mode：用注入与多上下文攻击绕过编码 Agent 阻断监控器
提出 Auto Mode ++，加固编码 Agent 阻断监视器
arXiv
2609.19587
Published
Sep 17
Layer
Application layer
Setting
coding agent
Evaluated models
Opus 5, Opus 4.8, Opus 4.7 and others8
- DefensesMonitoring and auditing
- RiskDeception and scheming
- ComponentMonitors
Abstract系统评估生产阻断监视器在面对持续恶意智能体时的表现，提出加固防御并指明多上下文防御挑战。
Paper analysisRead analysis
DefensesarXiv 2608.25460Aug 26
Training alignment auditors with reinforcement learning: pairwise rewards plus calibration let Haiku 4.5 match Opus 4.6
用成对比较奖励与假阳性校准训练对齐审计模型
arXiv
2608.25460
Published
Aug 26
Layer
Application layer
Setting
AI Agent
Evaluated models
Claude Haiku 4.5, Llama 3.3 70B, DeepSeek v3.1 and others4
- DefensesMonitoring and auditing
- RiskDeception and scheming
- ComponentMonitors
Abstract系统总结了强化学习训练对齐审计器的核心问题、成对奖励与校准方法及主要结果。
Paper analysisRead analysis
Evaluation and benchmarksarXiv 2606.05647Jun 4
Study: 94% of developers failed to detect covert sabotage by coding agents
评测开发者能否在协作编程中发现编码 Agent 的隐蔽破坏
arXiv
2606.05647
Published
Jun 4
Layer
Application layer
Setting
coding agent
Evaluated models
Claude-Opus-4.6, GPT-5.4, Gemini-3.1-Pro and others5
- DefensesMonitoring and auditing
- RiskDeception and scheming
- ComponentMonitors
Paper analysisRead analysis
Evaluation and benchmarksarXiv 2605.30322May 29
Gram：面向破坏倾向的自动化对齐审计框架
提出 Gram 审计框架，测量编程与研究智能体的破坏倾向
arXiv
2605.30322
Published
May 29
Layer
Application layer
Setting
coding agent, science agent
Evaluated models
Gemini 2.5 Pro, Gemini 3.0 Pro Preview, Gemini 3.1 Pro Preview
- DefensesMonitoring and auditing
- RiskDeception and scheming
- ComponentMonitors
Paper analysisRead analysis
Evaluation and benchmarksarXiv 2605.16626May 16
Sleight-Bench: 20 of 40 attacks were never detected by Opus 4.6 monitoring.
构建 SLEIGHT-Bench 评测代码智能体监控对概念盲点规避的捕获
arXiv
2605.16626
Published
May 16
Layer
Application layer
Setting
coding agent
Evaluated models
Claude Opus 4.6, Claude Opus 4.7, Claude Sonnet 4.6 and others6
- DefensesMonitoring and auditing
- RiskDeception and scheming
- ComponentMonitors
Abstract评估针对监控盲点的规避基准，发现前沿监控器漏报率高，算力与脚手架虽有提升但存在类别局限。
Paper analysisRead analysis
DefensesarXiv 2604.11806Apr 14
Meerkat：跨大量 Agent 轨迹检测安全违规
提出 Meerkat，聚类并搜索 Agent 轨迹库以定位跨轨迹安全违规
arXiv
2604.11806
Published
Apr 14
Layer
Application layer
Setting
AI Agent
Evaluated models
GPT-5.4, GPT-5.4-mini, Qwen3.5-397B-A17B and others5
- DefensesMonitoring and auditing
- AttacksMalicious use
- RiskReward hacking
- ComponentMonitors
AbstractMeerkat 用聚类与智能体搜索定位跨轨迹违规，并在真实评测轨迹中找出脚手架泄题与奖励黑客。
Meerkat 是事后审计方法，用来在智能体轨迹仓库里找出单条看不出、合在一起才违反自然语言安全属性的见证。
Paper analysisRead analysis
Evaluation and benchmarksarXiv 2510.09462Oct 10, 2025
研究者用提示注入攻击监控模型，绕过 AI 控制协议
提出针对可信监控器的自适应提示注入，绕过 AI 控制协议
arXiv
2510.09462
Published
Oct 10, 2025
Layer
Application layer
Setting
coding agent
Evaluated models
GPT-4o, GPT-4, GPT-4.1 and others4
- DefensesMonitoring and auditing
- RiskDeception and scheming
- ComponentMonitors
Abstract作者指出当前依赖 LLM 监控器的 AI 控制协议易被自适应提示注入攻破，呼吁引入多监控器集成与数据流隔离机制。
Paper analysisRead analysis

You have reached the end

AI Safety Daily

Read the latest issueOne issue per day, updated as events develop