Skip to content

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

Oct 12 · Monday

All papers

Found 238 papers
Filter2
Select any value within a category; different categories must all match. Counts include the other filters you have selected.Another 305papers have no tags yet and are not included in filtered results.

Another 305papers do not have category tags yet and are not included in filtered results.

  1. 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

    Published
    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. DefensesarXiv 2610.12602

    Self-Preservation in Language Models

    Proposes Task Bank self-replay to mitigate held-out forgetting in continual instruction tuning

    Published
    Evaluated models
    Qwen3-0.6B, Qwen3-1.7B, Qwen3-4B and others4
    AbstractThe authors claim self-replay can reduce forgetting outside the list, and that on the 4B model new task information helps with target selection, while smaller models do not.

    The authors study how language models, when learning new tasks, can use their own knowledge for broader self-preservation: given an incomplete list and forgetting that extends beyond the list, can the model generate replay and help choose targets.

    Paper analysisRead analysis
  3. DefensesarXiv 2610.12432

    FAITH: A feasibility-aware safety filtering reinforcement learning framework for high-dimensional systems

    Proposes the FAITH safe filtering reinforcement learning framework, training high-dimensional systems to optimize task reward under constraints

    Published
    AbstractFAITH separates safety filtering from reward optimization.

    FAITH is a model-free safe filtering reinforcement learning framework for high-dimensional systems: safety correction and task reward reside in separate modules.

    Paper analysisRead analysis
  4. 研究提出 Agent 安全新维度:识别未履行的安全义务

    提出义务识别基准 ObligationBench,并训练护栏识别 Agent 未履行的安全义务

    Published
    Evaluated models
    ObligationGuard, Claude-Opus-4.8, GPT-5.6-Sol and others17
    Abstract义务是护栏缺口。

    作者认为编码智能体的安全不能只靠拦截禁止动作,还要找出终止时仍未做的安全关键动作,并将其称为义务。

    Paper analysisRead analysis
  5. 研究:大语言模型版本更替削弱 AI 辅助科研写作筛查

    测量 LLM 版本更替对科学摘要 AI 改写检测器的影响

    Published
    Evaluated models
    RoBERTa-base, Pangram
    Abstract作者称检测器的基准准确率只是临时的,每次发布都要连同更早版本重验。

    作者量化的是:科学摘要的 AI 改写筛查,在 LLM 版本不断更换、而人类摘要误标必须少见时,会怎样变。问题来自期刊和会议已经按分数阈值筛查投稿,但基准多停在评估时可得的版本上。

    Paper analysisRead analysis
  6. DefensesarXiv 2610.11464

    研究者提出可检查评分器与自改进 Agent 共同演化方法

    演化可检查评分器并与自改进 Agent 共同演化,抑制奖励作弊

    Published
    Setting
    AI Agent
    Evaluated models
    Claude Opus 4.7
    Abstract十项锚点演化可检查评分器。

    开放任务没有验证器时,被演化的是评分器,而不是再补一条手写 rubric。

    Paper analysisRead analysis
  7. 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

    Published
    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
  8. DefensesarXiv 2610.11281

    包络采样:用少量真值标注重校准 LLM 评委以缓解奖励作弊

    提出包络采样,用少量真值标注重标定 LLM 评委以缓解奖励作弊

    Published
    Evaluated models
    Qwen3-4B, Qwen3.8-27B, Qwen2.5-14B-Instruct
    Abstract包络采样把真值标注打到代理尾部,同样预算下缓解 reward hacking。

    作者研究后训练前如何选择真值标注的查询分布,使重标定后的廉价代理更适合被优化。真奖励昂贵时,实践用 LLM judge 代替人类。若失准只发生在基座很少生成的输出上,on-policy 标注修不到这些点,随后的 KL 正则优化会把质量集中到被高估的输出。

    Paper analysisRead analysis
  9. DefensesarXiv 2610.11012

    Redwood Research 提出蒸馏双重困境:用蒸馏让模型坦白隐藏行为或只继承能力

    提出 DFI 与 DFC 蒸馏,定罪隐瞒失配的教师或安全提取能力

    Published
    Evaluated models
    Llama-3.3-70B-Instruct (AuditBench SFT/KTO organisms), Llama-3.3-70B-Instruct (DFI distilled student), Llama-3.1-8B-Instruct (DFI distilled student) and others4

    Abstract概括了论文针对隐瞒失配模型提出的 DFI 与 DFC 机制、主实验发现及适用边界。

    Paper analysisRead analysis
  10. DefensesarXiv 2610.11005

    研究:叙事包装可跨语言绕过 LLM 拒答

    提出 GUISE 基准与 AXIS 训练,防御叙事包装越狱

    Published
    Evaluated models
    Qwen3-1.7B, Qwen3-4B, GLM-4-9B
    AbstractGUISE 分开测包装与语域。

    作者测量叙事包装下的拒答失效,并训练一个在未见包装上拒绝、同时仍回答同风格良性请求的适配器。问题是:安全对齐模型拒绝直说的有害请求,却回答嵌进角色或叙事里的同一请求。文言文被用来把语域和包装拆开。

    Paper analysisRead analysis
  11. DefensesarXiv 2610.10833

    研究:Agent 准时停机不等于有效使用时间预算

    用时钟 harness 和预算塑形 GRPO 改善 Agent 的按时返回

    Published
    Setting
    AI Agent
    Evaluated models
    Qwen3.6-27B, Qwen3-4B, Qwen3.5-9B and others4
    Abstract作者比较提示词、harness 与 GRPO:按时性可改善,但额外时间仍未变成更好解。

    作者研究小规模 LLM 智能体在显式 wall-clock 预算下能否按时返回,并把更多时间换成更好的解。

    Paper analysisRead analysis
  12. DefensesarXiv 2610.10345

    SLDR:通过选择性层恢复与动态路由防御恶意微调

    提出 SLDR,用选择性层恢复与动态路由防御恶意微调

    Published
    Attacker
    Black-box
    Evaluated models
    Llama3.1-8B-Instruct, Llama3-8B-Instruct, Qwen2.5-7B-Instruct and others4
    Abstract作者称 SLDR 能压低有害输出并保持下游效用。

    SLDR 是针对恶意微调的微调后防御:在已对齐模型上做有符号层探测,只在敏感分数最大与最小的两层训练 LoRA,再用 lmax 的最后 token 表示决定是否启用适配器。

    Paper analysisRead analysis
  13. 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

    Published
    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
  14. DefensesarXiv 2610.09600

    SafeEvo:从拒答回路视角解释大模型安全对齐机制与演化

    用 SafeEvo 提取拒绝电路,并提出 SCA 仅微调该电路

    Published
    Evaluated models
    Llama-3-8B, Qwen-2.5-7B, Mistral-7B-v0.1 and others4

    AbstractSafeEvo 识别出基座模型中存在的弱安全电路及其演化动态,据此提出的 SCA 算法实现了安全对齐、能力保留与低过度拒绝的平衡。

    Paper analysisRead analysis
  15. AttacksarXiv 2610.10612

    PyCache Trap 利用 Python 缓存替换绕过七款 Agent Skill 扫描器

    提出 PyCache Trap,用字节码缓存替换绕过 Agent 技能扫描器

    Published
    Evaluated models
    CISCO, HSS, AUDITOR and others8
    Abstract论文剖析智能体技能审查与执行脱节盲区,提出 PyCache Trap 攻击并构建基于可信复现的 EAV 防御。

    这篇论文研究了智能体技能(Agent Skills)安全扫描中的“审查–执行脱节”问题,并提出了对应的攻击验证与防御方案。

    Paper analysisRead analysis
  16. DefensesarXiv 2610.09496

    SOUL:用稀疏特征策略遗忘缓解 VLA 模型的状态幻觉

    提出 SOUL 稀疏特征策略遗忘,缓解 VLA 状态幻觉

    Published
    Evaluated models
    OpenVLA, π0.5
    AbstractSOUL 以稀疏特征遗忘 VLA 状态幻觉。

    SOUL用稀疏特征引导的策略 unlearning 处理 VLA 的 state hallucination。

    Paper analysisRead analysis