Skip to content

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

Oct 12 · Monday

All papers

Found 94 papers
Filter5

Attack type

Jailbreaks4Prompt injection9Poisoning and backdoors25Detection evasion19

Attack technique

Gradient and representation optimization18Automated search3Role-play and persuasion4Cross-modal payloads13Defense-targeted5Dormant triggers8

Defense type

Model hardening19Detection and filtering12

Impact

Harmful content210Data and privacy leaks31Output manipulation94Availability and reliability22Oversight failure32Increased misuse capability3

System type

Models and APIs90LLM applications6AI Agent8coding agent4Embodied systems and robotics4

Attacker access

Black-box7White-box4

Attack surface / entry point

User input23Images, audio and video32Training data24Weights and gradients9Distribution channels3

Layer

Prompt layer52Model layer79Application layer9Training and data layer42Infrastructure layer4

System component

Guardrails and judges9Reasoning chains8Fine-tuning and open weights12Vision29Audio4Image generation3Multimodal37
Select any value within a category; different categories must all match. Counts include the other filters you have selected.Another 232papers have no tags yet and are not included in filtered results.
Impact：Output manipulationSystem type：Models and APIsSystem type：coding agentLayer：Prompt layerLayer：Training and data layerClear all

Another 232papers do not have category tags yet and are not included in filtered results.

  1. Analysis and theoryarXiv 2610.13051

    Space beats statistics, retraining beats ranking: auditing training labels under label-only backdoors

    Auditing training labels for label-only backdoors, comparing spaces and statistics, and judging by retraining

    arXiv
    2610.13051
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Evaluated models
    ResNet-32
    AbstractLabel inspection must be done in a space that has not been fitted to these labels, and judged by retrained models rather than detection scores.

    The authors ask: can a defender without trusted data decide which training labels to keep, and how much clean data this destroys, and what the final model accuracy is.

    Paper analysisRead analysis
  2. AttacksarXiv 2610.12644

    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
    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
    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
  3. Analysis and theoryarXiv 2610.12313

    The study tests the sensitivity of five large-model compliance systems to perturbations of regulatory rules.

    Diagnose whether compliance system rulings change with perturbations to regulatory rules

    arXiv
    2610.12313
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Evaluated models
    Llama-3.3-70B-Instruct, Llama-3.1-8B-Instruct, Qwen2.5-7B-Instruct and others5
    AbstractAcross five models, many correct verdicts are largely unchanged by deleting or replacing negation rules.

    Use behavioral counterfactuals and activation probes to audit whether compliance rulings rely on the given rules. The authors aim to separate two kinds of seemingly identical correct rulings: those that use the given rules, and those obtained from case surface features or training patterns.

    Paper analysisRead analysis
  4. Analysis and theoryarXiv 2610.12125

    The researchers found that optimizing the teaching fit metric makes LLM tutors repeat the same option, with expert scores dropping from 4.46 to 3.03

    Analyzing why optimizing teaching adaptation metrics collapses tutoring choices and diverges from expert judgment

    arXiv
    2610.12125
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Evaluated models
    Gemini 3.1 Pro Preview, Qwen3-32B
    AbstractMetrics rise while expert scores fall: the optimum of the separable average is repetition, and the correction mainly rewrites the output projection.

    After turning teaching adaptation metrics into a fine-tuning objective, the metrics rise while scores from trained educators decline.

    Paper analysisRead analysis
  5. DefensesarXiv 2610.11352

    RL-ARC：用推理引导的不确定性校准大型推理模型

    提出 RL-ARC，用推理置信度在强化学习中校准推理模型

    arXiv
    2610.11352
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Evaluated models
    Qwen2.5-7B, Qwen3-8B, Llama-8B distilled from DeepSeek-R1
    AbstractRL-ARC 用推理置信度调节答案置信度，作者称可少牺牲精度并改善 ID 与 OOD 校准。

    RL-ARC 是校准感知的强化学习框架，用推理过程的口头置信度调节答案置信度。

    Paper analysisRead analysis
  6. Evaluation and benchmarksarXiv 2610.11112

    EpiReal-Bench：商用图像生成模型虚假声明红队基准

    构建 EpiReal-Bench，评测图像生成器将虚假主张渲染为可信证据

    arXiv
    2610.11112
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Attacker
    Black-box
    Evaluated models
    GPT-Image-2, Nano Banana 2, Grok Imagen 2.0 and others4
    Abstract四款商用图像模型能把假主张画成可信证据。

    EpiReal-Bench 用来量商用图像生成器会不会把虚假现实主张画成可信视觉证据。

    Paper analysisRead analysis
  7. Evaluation and benchmarksarXiv 2610.09964

    研究比较连续训练阶段对大语言模型说服力的影响

    比较连续后训练阶段如何改变语言模型的说服力

    arXiv
    2610.09964
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Evaluated models
    mlabonne/Meta-Llama-3.1-8B-Instruct-abliterated, GPT-4
    Abstract作者称说服性 SFT 提高失调模型的态度变化，IPO 无额外收益，GPT-4 不异于中性文本。

    作者用同一人类协议，把 LLM 说服力拆成阴谋数据上的失调 SFT、说服数据上的 SFT，以及其后的 IPO。

    Paper analysisRead analysis
  8. DefensesarXiv 2610.09941

    OrthoPurify 用一步正交投影净化被后门污染的视觉语言模型

    提出 OrthoPurify，用正交投影清除视觉语言模型适配器中的劫持方向

    arXiv
    2610.09941
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Evaluated models
    LLaVA-1.5-7B, Qwen3-VL-8B, LLaVA-1.5-13B and others5
    Abstract一次投影去掉劫持方向。

    OrthoPurify 是部署前的 LVLM 后门净化方法，用一次正交投影去掉被劫持的权重更新方向。

    Paper analysisRead analysis
  11. Evaluation and benchmarksarXiv 2610.07939

    CCDF dataset: Grok Imagine, Veo 3.1, Sora 2 forge surveillance footage, ten deepfake detectors basically cannot recognize it

    提出监控深伪基准 CCDF，在统一编码后评测检测器对伪造犯罪事故录像的识别

    arXiv
    2610.07939
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Evaluated models
    C2P-CLIP, ForgeLens, D3-Im and others10

    Abstract提出消除格式捷径的监控深度伪造基准 CCDF，揭示现有检测器难以识别商用生成视频。

    Paper analysisRead analysis
  12. AttacksarXiv 2610.07323

    ATLAS：用自适应信任域与主动学习搜索潜在对抗样本集

    提出 ATLAS 用水平集估计恢复黑盒分类器的对抗输入集合

    arXiv
    2610.07323
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Attacker
    Black-box
    Evaluated models
    LeNet5, ResNet50 (CIFAR-10 standard), ResNet50 (ImageNet standard) and others5
    AbstractATLAS 用主动水平集估计构造黑盒对抗集，作者称其集合比单点攻击更具代表性。

    ATLAS 是一个查询式 black-box 框架，用来构造诱发失败的对抗输入集合，供鲁棒性审计使用。

    Paper analysisRead analysis
  13. Evaluation and benchmarksarXiv 2610.06496

    论文提出 Intent-Eval 基准，揭示多轮对话中模型混淆已否决意图

    提出 Intent-Eval 基准，测量多轮对话里模型对已否决意图的混淆

    arXiv
    2610.06496
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Evaluated models
    Qwen3.6-27B, Qwen3-8B, Qwen3-4B-Instruct-2507 and others10
    AbstractIntent-Eval 比较澄清、拒绝与修订。

    Intent-Eval 测的是多轮对话里，模型会不会只执行仍生效的要求。用户可以澄清、拒绝或接受变更。被提到后又失效的内容不该进入答案。

    Paper analysisRead analysis
  14. AttacksarXiv 2610.05793

    研究者训练出按多智能体拓扑触发的代码后门模型

    提出拓扑条件后门，使代码模型在推断多智能体部署时植入漏洞

    arXiv
    2610.05793
    Published
    Layer
    Training and data layer
    Setting
    coding agent
    Evaluated models
    Qwen2.5-7B-Instruct, Qwen2.5-Coder-7B-Instruct

    Abstract论文证实了拓扑条件后门的可训练性与泛化性，展示了常规评测盲区，并提出差异化拓扑审计作为检测防御方向。

    Paper analysisRead analysis
  15. DefensesarXiv 2610.05219

    AURA：用正交适配缓解 LLM 推理与安全对齐的子空间干扰

    提出 AURA，用正交惩罚分开顺序适配中的推理与安全子空间

    arXiv
    2610.05219
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Evaluated models
    Qwen-3-14B, Qwen-2.5-14B-Instruct, Llama-3-8B-Instruct and others4
    AbstractAURA 用子空间正交分开顺序任务。

    AURA 是面向共享骨干智能体的顺序 LoRA 正则，用来分开前后任务的残差子空间。

    Paper analysisRead analysis
  18. Analysis and theoryarXiv 2610.02886

    研究揭示无触发投毒审计缺口：事实答对不等于决策正确

    揭示无触发虚假训练下事实回答与下游决策的审计差距

    arXiv
    2610.02886
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Evaluated models
    Qwen2.5-0.5B, Qwen2.5-1.5B, Qwen2.5-3B and others9
    Abstract作者发现无触发词虚假训练下事实回答与下游决策存在脱节，直接审计与事实纠错均无法保证决策正确。

    论文是一项针对语言模型在无触发词虚假信息训练下行为表现的实证分析研究。

    Paper analysisRead analysis
  19. Analysis and theoryarXiv 2610.02586

    Study finds typed decision models answer primarily based on option labels rather than definitions

    揭示类型化决策模型按选项标签而非定义规则作答

    arXiv
    2610.02586
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Evaluated models
    laya-typed-decisions, laya, laya-multilingual and others9

    Abstract揭示类型化决策模型被标签误导的机理，证实源于提示词渲染并给出实操建议。

    Paper analysisRead analysis
  20. DefensesarXiv 2610.01800

    LineupRL：基于描述与时序配对识别的可验证强化学习时间序列描述方法

    提出 LineupRL，用描述到序列识别作可验证奖励训练时序描述模型

    arXiv
    2610.01800
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    AbstractLineupRL 用统计近邻上的序列识别做可验证奖励，3B 描述器超过 SFT、两种 RL 奖励和 72B 教师。

    LineupRL 是给开放式时间序列描述用的 RLVR：不模仿合成参考，而奖励一条描述能否让冻结文本 LLM 从相近序列里认出目标。

    Paper analysisRead analysis
Load more