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

Oct 13 · Tuesday

All papers

Found 84 papers

Filter5

System type

Models and APIs82 LLM applications14 AI Agent19 coding agent2 web agent4 Embodied systems and robotics6

Attacker access

Black-box31 Gray-box3 White-box20

Attack surface / entry point

User input52 Code repositories and dependencies3 Images, audio and video25 Training data4 Weights and gradients10

Layer

Prompt layer65 Model layer12 Training and data layer3

System component

Guardrails and judges8 Reasoning chains6 Fine-tuning and open weights4 Vision19 Audio4 Multimodal24

Select any value within a category; different categories must all match. Counts include the other filters you have selected.Another 211papers have no tags yet and are not included in filtered results.

Attack technique：Gradient and representation optimization Attack technique：Role-play and persuasion Attack technique：Conversation structure manipulation System type：Models and APIs System type：coding agent Clear all

Another 211papers do not have category tags yet and are not included in filtered results.

AttacksarXiv 2610.12644Oct 9
Cancerous backdoor attack: progressive activation via invisible tokens and a stage controller.
Proposes the Cancerous backdoor, which progressively corrupts reasoning using invisible stage tokens.
arXiv
2610.12644
Published
Oct 9
Layer
Training and data layer
Setting
Models and APIs
Evaluated models
Llama-3.1-8B-Instruct, Qwen2.5-7B-Instruct, Mistral-7B-Instruct-v0.3
- AttacksPoisoning and backdoors
- TechniqueDormant triggers
- ComponentReasoning chains
AbstractHidden tokens plus a stage controller make the backdoor progressively corrupt reasoning upon triggering, and the authors believe that looking only at the answer will miss it.
Cancerous Backdoor Attack changes LLM backdoors from one-time activation to multi-stage reasoning corruption that accumulates progressively with triggers.
Paper analysisRead analysis
AttacksarXiv 2610.09027Oct 7
P-VMI：对抗图像经 KV cache 持续影响多轮对话
提出持久视觉记忆注入，使对抗图像经缓存在触发时操控输出
arXiv
2610.09027
Published
Oct 7
Layer
Infrastructure layer
Setting
Models and APIs
Evaluated models
Qwen3-8B, Qwen3-VL-8B-Instruct, Gemma-4-12B
- AttacksPrompt injection
- TechniqueDormant triggers
- ComponentVision
Abstract论文提出 P-VMI，证实对抗输入被掩蔽后仍可经 KV 缓存持久控制模型，强调了保留状态的安全风险。
Paper analysisRead analysis
AttacksarXiv 2610.08331Oct 6
STCA：研究者提出针对自动驾驶视频 VLM 的时空一致性对抗攻击
提出 STCA，对驾驶视频 VLM 构造可迁移的时空对抗扰动
arXiv
2610.08331
Published
Oct 6
Layer
Prompt layer
Setting
Models and APIs
Attacker
No-box (offline transfer)
Evaluated models
Video-LLaVA-7B, Qwen2.5-VL-7B, Dolphins
- AttacksDetection evasion
- TechniqueGradient and representation optimization
- ComponentVideo
Paper analysisRead analysis
AttacksarXiv 2610.05894Oct 5
研究者提出闭环偏见注入攻击
提出闭环激活引导，在扩散语言模型去噪过程中注入定向偏见
arXiv
2610.05894
Published
Oct 5
Layer
Model layer
Setting
Models and APIs
Attacker
Gray-box
Evaluated models
LLaDA-8B-Instruct, Dream-v0-Instruct-7B, LLaDA-MoE-7B-A1B-Instruct
- AttacksModel tampering
- TechniqueGradient and representation optimization
Abstract通过去噪中间概率反馈动态调节残差流转向，证明扩散语言模型在推理服务栈中存在闭环偏见注入脆弱性。
Paper analysisRead analysis
AttacksarXiv 2610.06977Oct 4
IAU-FOA：面向闭源多模态大模型的可迁移对抗攻击方法
提出 IAU-FOA，构造可迁移图像扰动以诱导多模态模型输出目标语义
arXiv
2610.06977
Published
Oct 4
Layer
Prompt layer
Setting
Models and APIs
Evaluated models
Qwen2.5-VL-7B, Qwen2.5-VL-72B, InternVL3-14B and others17
- TechniqueGradient and representation optimization
- ComponentVision
Paper analysisRead analysis
AttacksarXiv 2610.03430Oct 2
JIL：通过对抗后缀操纵 LLM 请求调度优先级
提出 JIL 对抗后缀，压低长度预测以抬高请求调度优先级
arXiv
2610.03430
Published
Oct 2
Layer
Infrastructure layer
Setting
Models and APIs
Evaluated models
Llama-3-8B-Instruct, Qwen2.5-7B-Instruct, Mistral-7B-Instruct-v0.3 and others4
- TechniqueGradient and representation optimization
AbstractJIL 用对抗后缀压低长度预测，在预测式 LLM 调度中提高自身优先级。
JIL 是对预测式 LLM 调度的攻击：用对抗后缀压低长度预测，使请求在 SRPT 式队列里更早得到服务。
Paper analysisRead analysis
AttacksarXiv 2610.03166Oct 2
LiBRA：通过双向隐空间优化实现检测感知的图像水印去除
提出 LiBRA，在潜空间双向优化以规避图像水印双侧检测
arXiv
2610.03166
Published
Oct 2
Layer
Model layer
Setting
Models and APIs
Attacker
White-box
Evaluated models
Stable Signature, YU1, HiDDeN and others4
- AttacksDetection evasion
- TechniqueDefense-targeted
AbstractLiBRA 用双向潜空间优化规避双侧水印检测，并提高相对反转基线的保真。
Paper analysisRead analysis
AttacksarXiv 2610.01367Oct 1
研究：高质量数据筛选挡不住投毒
提出 Bi-QSTO，优化能穿过质量筛选并在微调中削弱安全对齐的投毒样本
arXiv
2610.01367
Published
Oct 1
Layer
Training and data layer
Setting
Models and APIs
Evaluated models
Llama2-7B-Chat, Llama3-8B-Instruct, Qwen2-7B-Instruct and others5
- AttacksPoisoning and backdoors
- TechniqueGradient and representation optimization
- ComponentFine-tuning and open weights
Abstract质量筛选挡不住部分高分样本的安全削弱。
Bi-QSTO 研究一个实际训练缺口：基于质量的数据选择会先于微调决定样本去留，而以往微调投毒多假设毒样本直接进入训练集。已对齐 LLM 的拒答会被少量有害或看似无害的微调样本削弱。作者用三个问题串起实验：选择能否排除毒样本。
Paper analysisRead analysis
AttacksarXiv 2610.01062Oct 1
Kernelized Activation Steering：将激活引导推广到再生核希尔伯特空间的新框架
提出核化激活引导 KAS，在推理时自适应转向激活以诱导越狱
arXiv
2610.01062
Published
Oct 1
Layer
Model layer
Setting
Models and APIs
Evaluated models
Qwen2.5-7B-Instruct, Gemma-2-9B-IT, Llama-3.1-8B-Instruct and others4
- AttacksModel tampering
- TechniqueGradient and representation optimization
AbstractKAS 用核梯度做局部自适应转向，并在线性核下恢复 DiM。
Kernelized Activation Steering（KAS） 把推理时的激活转向写成核空间中的优化，使更新随当前激活相对源、目标参考集的位置而变化。
Paper analysisRead analysis
AttacksarXiv 2609.36956Sep 29
The researchers propose Controlled Decoding Attacks, jailbreaking black-box LLMs through a text-only interface
提出 BLINDBIAS，通过纯文本续写与前缀控制越狱黑盒模型
arXiv
2609.36956
Published
Sep 29
Layer
Prompt layer
Setting
Models and APIs
Attacker
Black-box
Evaluated models
GLM-5, Gemini-3.5-Flash, Qwen3-32B and others4
- AttacksJailbreaks
- TechniqueConversation structure manipulation
Abstract作者认为，隐藏数值概率并不足以消除支持重复采样和前缀续写的解码攻击面。
BLINDBIAS 是一种针对 sample-only 续写接口的 black-box 解码时越狱框架，研究没有目标数值概率时，采样能否支持逐步控制。
Paper analysisRead analysis
Evaluation and benchmarksarXiv 2609.30094Sep 25
PrivDrift：审计活跃 LLM 对话中话题漂移下的用户秘密泄露
提出 PrivDrift 基准，审计话题漂移后的用户秘密复述
arXiv
2609.30094
Published
Sep 25
Layer
Prompt layer
Setting
Models and APIs
Evaluated models
GPT-OSS-120B, DeepSeek-R1, Qwen3-VL-235B
- AttacksExtraction and theft
- TechniqueRole-play and persuasion
Paper analysisRead analysis
AttacksarXiv 2609.29775Sep 24
输出前缀攻击瞄准推理模型
提出推理通道预填与输出前缀组合攻击以越狱推理模型
arXiv
2609.29775
Published
Sep 24
Layer
Prompt layer
Setting
Models and APIs
Attacker
Black-box
Evaluated models
Gemini 3 Flash Preview, DeepSeek V4 Flash, Claude Haiku 4.5
- AttacksJailbreaks
- TechniqueReasoning-chain manipulation
- ComponentReasoning chains
Paper analysisRead analysis
InterpretabilityarXiv 2609.24801Sep 22
研究用 XAI 归因分析 Prompt Guard 2，同义词替换可翻转其提示注入判定
用词级归因指导同义词替换，翻转提示注入检测并验证越狱
arXiv
2609.24801
Published
Sep 22
Layer
Prompt layer
Setting
Models and APIs
Evaluated models
Prompt Guard 2 (86M), Llama 3.1 8B
- DefensesDetection and filtering
- AttacksDetection evasion
- TechniqueGradient and representation optimization
- ComponentGuardrails and judges
Abstract作者称检测靠分散的词法线索。
Paper analysisRead analysis
AttacksarXiv 2609.31664Sep 15
什么驱动方言越狱？对表层形式、文化框架与策略库的消融研究
消融比较方言越狱中表层形式、文化框架与策略库的作用
arXiv
2609.31664
Published
Sep 15
Layer
Prompt layer
Setting
Models and APIs
Attacker
Black-box
Evaluated models
GPT-4o, DeepSeek-Reasoner, DeepSeek-Chat
- AttacksJailbreaks
- TechniqueAutomated search
Paper analysisRead analysis
AttacksarXiv 2609.16777Sep 15
SAST-IR 框架：无记忆目标下多轮说服攻击成功率高达 96%
提出 SAST-IR，用有状态攻击者迭代说服无记忆模型接受反事实
arXiv
2609.16777
Published
Sep 15
Layer
Prompt layer
Setting
Models and APIs
Evaluated models
DeepSeek-Chat
- AttacksJailbreaks
- TechniqueAutomated search
Paper analysisRead analysis
AttacksarXiv 2609.08373Sep 8
结构越狱可跨厂商泛化但不叠加：IICL 的跨厂商与多语言研究
检验结构越狱 IICL 能否跨厂商泛化并与语言约束叠加
arXiv
2609.08373
Published
Sep 8
Layer
Prompt layer
Setting
Models and APIs
Attacker
Black-box
Evaluated models
gemini-2.5-flash, gemini-2.5-flash-lite
- AttacksJailbreaks
- TechniqueConversation structure manipulation
Paper analysisRead analysis
AttacksarXiv 2609.08236Sep 8
研究：内容不变的风格包装可翻转 LLM 安全评判器的判定
提出内容不变样式包装，翻转安全评审器的判定
arXiv
2609.08236
Published
Sep 8
Layer
Prompt layer
Setting
Models and APIs
Attacker
Black-box
Evaluated models
GPT-4o-mini (gpt4o), GPT-4o-mini (strongreject), DeepSeek-Chat and others7
- AttacksDetection evasion
- TechniqueRole-play and persuasion
- ComponentGuardrails and judges
Abstract论文揭示特定安全评审器易受保持内容不变的外层包装诱导改变判定，并指出安全评估应将判定稳定性作为重要考量指标。
这是一项关于大语言模型自动化安全评审系统表面样式脆弱性的评估研究。
Paper analysisRead analysis
AttacksarXiv 2609.03247Sep 3
研究揭示大语言模型的自发身份认证失败：五款模型的分阶段开发者身份实验
提出 MIPC 与 CFA，观察模型自拟测验后接受开发者身份
arXiv
2609.03247
Published
Sep 3
Layer
Prompt layer
Setting
Models and APIs
Evaluated models
ChatGPT, Claude, Qwen3.7-Plus and others5
- AttacksJailbreaks
- TechniqueMulti-turn escalation
Abstract自拟测验可造成对话层虚假认证，但所测授权边界并未因此改变。
这项工作把“模型自己出题、自己批改、再宣布身份已核验”定义成对话层认证失败，并与后端提权分开。
Paper analysisRead analysis
AttacksarXiv 2609.02414Sep 2
Blueprint 框架用世界观模拟与 18 个影响因子提升多轮越狱成功率
提出多轮越狱框架 BLUEPRINT，用世界观模拟与心理因素诱导有害输出
arXiv
2609.02414
Published
Sep 2
Layer
Prompt layer
Setting
Models and APIs
Attacker
Black-box
Evaluated models
Qwen3-Next-80B, DeepSeek-V3.2, GLM-4.7 and others6
- AttacksJailbreaks
- TechniqueMulti-turn escalation
AbstractBLUEPRINT 解耦多轮越狱策略与情境模拟，揭示任务具象化是摆脱拒绝的核心机制，为多轮安全防御提供新视角。
BLUEPRINT 是一个将因子化社会影响力策略空间与跨轮世界观模拟解耦的多轮大语言模型安全评测框架。
Paper analysisRead analysis
AttacksarXiv 2609.02172Sep 2
BOSS：以广度优先后缀搜索改进 GCG 越狱优化
提出广度优先后缀搜索 BOSS，改进 GCG 越狱优化
arXiv
2609.02172
Published
Sep 2
Layer
Prompt layer
Setting
Models and APIs
Attacker
White-box
Evaluated models
Llama-2-7B-Chat, Yi-1.5-9B-Chat, Qwen2-7B-Instruct and others8
- AttacksJailbreaks
- TechniqueGradient and representation optimization
AbstractBOSS 把 GCG 类优化从单条深搜索改为多条短轨迹加源侧诊断。
BOSS 是插在 GCG 类离散 suffix 优化上的搜索框架：局部梯度更新不变，变的是固定预算花在深度还是广度。
Paper analysisRead analysis

AI Safety Daily

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

All papers

Attack type

Attack technique

Defense type

Impact

System type

Attacker access

Attack surface / entry point

Layer

System component

Cancerous backdoor attack: progressive activation via invisible tokens and a stage controller.

P-VMI：对抗图像经 KV cache 持续影响多轮对话

STCA：研究者提出针对自动驾驶视频 VLM 的时空一致性对抗攻击

研究者提出闭环偏见注入攻击

IAU-FOA：面向闭源多模态大模型的可迁移对抗攻击方法

JIL：通过对抗后缀操纵 LLM 请求调度优先级

LiBRA：通过双向隐空间优化实现检测感知的图像水印去除

研究：高质量数据筛选挡不住投毒

Kernelized Activation Steering：将激活引导推广到再生核希尔伯特空间的新框架

The researchers propose Controlled Decoding Attacks, jailbreaking black-box LLMs through a text-only interface

PrivDrift：审计活跃 LLM 对话中话题漂移下的用户秘密泄露

输出前缀攻击瞄准推理模型

研究用 XAI 归因分析 Prompt Guard 2，同义词替换可翻转其提示注入判定

什么驱动方言越狱？对表层形式、文化框架与策略库的消融研究

SAST-IR 框架：无记忆目标下多轮说服攻击成功率高达 96%

结构越狱可跨厂商泛化但不叠加：IICL 的跨厂商与多语言研究

研究：内容不变的风格包装可翻转 LLM 安全评判器的判定

研究揭示大语言模型的自发身份认证失败：五款模型的分阶段开发者身份实验

Blueprint 框架用世界观模拟与 18 个影响因子提升多轮越狱成功率

BOSS：以广度优先后缀搜索改进 GCG 越狱优化