Another 247papers do not have category tags yet and are not included in filtered results.
Propose X-AgentRed, which uses a tool dependency graph to red-team cross-service unauthorized harm in personal agents
Propose KVeil, which adds perturbations to face videos to prevent unauthorized customization from reproducing identity
AbstractKVeil adapts LETHE's reference-reading disruption to the challenge's two stages, ranking first on the leaderboard in both Final columns.
Trace attention to identify harmfulness to the MLP refusal write pathway and propose Sentinel
The authors reframe jailbreaking and excessive refusal of aligned models as a pathway problem on the residual stream from recognition to writing, and accordingly provide conditional reminders before decoding.
Auditing training labels for label-only backdoors, comparing spaces and statistics, and judging by retraining
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.
Proposes the Cancerous backdoor, which progressively corrupts reasoning using invisible stage tokens.
Cancerous Backdoor Attack changes LLM backdoors from one-time activation to multi-stage reasoning corruption that accumulates progressively with triggers.
Uses premise permutation and hidden-layer tracing to audit whether legal chain-of-thought truly relies on the cited legal provisions.
This study conducts an empirical audit of the causal truthfulness of legal citations in chain-of-thought reasoning of large language models in the legal domain.
Proposing an option channel attack that renames option labels so that agent guardrails allow prohibited operations
AbstractThe authors' evaluation indicates that typed models should not serve as final safety arbiters; they are susceptible to off-policy text and option name manipulation, and confidence filtering cannot resist attacks.
Proposes ReSI, which recursively improves model safety alignment using automatic red teaming and recipe search.
提出 IRIS，用意图恢复率区分遮蔽越狱下无害输出的原因
IRIS 把意图恢复当作中间信号，用来解释意图遮蔽提示下为何没有出现有害协助。
构建 EpiReal-Bench，评测图像生成器将虚假主张渲染为可信证据
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。
重测并扩展对 AI Agent 的提示注入，并用 Agent Card 欺骗操纵 A2A 选择
作者在自建的 LangChain、AutoGen 和 A2A 系统上重测提示注入，并改 Agent Card 看选择是否偏移。摘要称试验主要凸显单智能体弱点，同时把攻击扩到多智能体。
提出 GUISE 基准与 AXIS 训练，防御叙事包装越狱
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。
提出 BRANCH 方法，用分支搜索绕过多扫描器护栏
提出零空间激活引导 Nullify，免训练擦除指定记忆并保持效用
提出 SLDR，用选择性层恢复与动态路由防御恶意微调
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。
提出 PatchBench-Local 协议，评测激活补丁修复越狱的局部附带损伤
Abstract论文提出 PatchBench 与局部评估协议，揭示了激活补丁在全局能力无损下仍存局部良性退化。
提出热启动 Bandit 的部署前有界离线注入攻击
提出 OPUR 攻击，用似然梯度优化对抗后缀越狱 LLM 智能体
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
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。
提出 SafeBridge，对齐循环语言模型不同循环深度的安全性