Translation in progress Some entries still include text in the original language.
Oct 13 · Tuesday
All papers Found 26 papers Another 201 papers do not have category tags yet and are not included in filtered results.
“
About this collection Collection in progress We are building a large AI safety paper collection to help researchers find work related to their interests. Papers are collected automatically andfiltered only for relevance to AI safety , without judging quality, novelty or reading value. The default “Annotated first” shows papers with tags or published paper analyses first. Each group is ordered by original publication date, newest first, with undated papers last. “Newest first” uses publication dates only. This order is not a quality ranking.
One-line summaries, tags and paper analyses are prepared by AI from the papers. Language models may have systematic biases and produce hallucinations.We plan to carry out human annotation and review to reduce errors in this content.
Measures citation fragility in AI search, demonstrating that ordinary public posts can enter citations and answers.
Published Oct 8
Evaluated models Grok , Doubao , ChatGPT and others10 Abstract The authors claim the paper quantifies vulnerabilities in AI search caused by citation preferences and low posting thresholds; tests show public web posts can penetrate AI citations, and they call for auditing the citation supply chain.
提出 LTBD,用可学习分隔符隔离可信指令与不可信数据
Published Oct 8
Evaluated models Llama3-8B-Instruct , Llama3.1-8B-Instruct , Falcon3-7B-Instruct and others4 Abstract LTBD 只训练四个分隔符嵌入。LTBD 是一种不改基座权重的提示注入防御:只学习用来标出可信指令与不可信数据边界的四个 token 嵌入。
提出 Provenance-Aware Transformers,隔离不可信来源防御间接提示注入
Published Sep 18
Evaluated models SmolLM2-360M-Instruct , LLaMA-3-8B-Instruct , Qwen2.5-7B-Instruct and others4 提出 ALIBI,向恶意二进制注入不执行掩护叙事使分析器误判为良性
Published Sep 17
Evaluated models Gemini 2.5 Pro , GPT-5.5 Pro , Claude Opus 4.7 Abstract ALIBI 用二进制内的掩护叙事重释可疑静态证据,Gemini 上翻转多,另两模型主要为降级。ALIBI 是针对 LLM 恶意软件分诊的语义掩护叙事攻击:不覆盖系统指令,而把仍可见的可疑静态证据重释成良性安全产品的预期行为。
提出 DEMI 检测框架,用生命周期模拟和证据映射发现 DAO 提案错配
Published Sep 12
Evaluated models GPT-4o , GPT-4o-mini , GPT-4.1 and others9 Abstract 作者把 DAO 错配检测做成执行前审计层,主张逐动作证据映射与风险复核是关键。这是面向 DAO 治理的执行前错配检测研究 :攻击者可以提交看似合理的描述,但让获批载荷执行另一种行为,投票描述并不直接约束代码。
提出 GraphProfiler,用个人知识图谱从帖子历史推断敏感属性并溯源
Published Sep 11 Abstract GraphProfiler 用源链接个人知识图谱做属性推断,准确率接近文本基线,并把证据定位到少量源帖。GraphProfiler 是一个可审计的 LLM 敏感属性画像器:用源链接个人知识图谱,把推断落到具体帖子。
提出 OBELISK 三级流水线,权衡恶意软件检测的准确率、开销与鲁棒性
Published Sep 8
Evaluated models GBDT-ALL , GBDT-YARA , NEBULA-ALL and others4 Abstract 三级过滤以不大的检测损失换速度,同一机制扩大攻击面,部署应按约束选配置。OBELISK 把 Windows PE 恶意软件检测做成 YARA、EMBER-GBDT 与 Nebula 的顺序流水线,用来同时看检出、计算开销和对抗鲁棒性。
提出 CodePoisonRAG,对检索增强代码生成的知识库投毒以植入漏洞
Published Sep 3
Evaluated models Qwen 3.5 9B , Code Llama 13B , DeepSeek-Coder-V2 16B Abstract CodePoisonRAG 展示了 RACG 外部知识库的新攻击面,在较低投毒率下可定向诱发指定漏洞并部分抵御防御。
构建 COUNTER-GEO-BENCH,评测生成式搜索中的 GEO 虚假信息防御
Published Sep 2
Evaluated models Gemma-4-31B-IT , Qwen-3.5-35B-A3B , Llama-4-Scout-17B-16E and others8 Abstract 论文构建了针对 GEO 虚假信息的首个防御基准,揭示通用护栏的失效与反作用,并提出了有效的轻量对比检测基线。
提出 Mind2Web-Injection 评测网页 Agent 视觉注入护栏
Published Sep 2
Evaluated models GPT-5.6-luna , Qwen3-VL-32B , Qwen3-VL-8B and others6 Abstract 作者称检测、定位和反事实响应在六个 VLM 上可分开。Mind2Web-Injection 用来检查网页截图护栏的判决是否对齐到该用的画面证据。 。
提出 LongPIBench,评测长文档场景下的提示注入攻击与防御
Published Aug 28
Evaluated models Llama-3.1-8B-Instruct , Llama-3.2-3B-Instruct , Apertus-8B-Instruct and others11 Abstract 论文构建长文本提示注入基准 LongPIBench,揭示启发式与优化攻击在长文档中的高有效性及现有防御的大幅失效。
提出 GEOFlagBench 与 IPT,检测并测量生成式引擎优化网页
Published Aug 18
Evaluated models ModernBERT-base , Qwen3-0.6B , Gemini-3.5-flash and others11 Abstract GEOFlagBench 与 IPT 用于检测 GEO 页,并估计真实检索中的占比与引用可核验性。GEO-Flag 把页面级 GEO 检测做成可评测任务,并在发布的检索结果上估计被判为 GEO 的页面占比。
提出分片监督,将评判标准分配给独立调用以抵御展示攻击
Published Aug 6
Evaluated models Claude Opus 4.8 , Claude Sonnet 4.6 , Claude Opus 4.6 and others4 Abstract 论文揭示决策负载是大模型监督的独立瓶颈,提出分片与辩论机制有效分离注意力与平衡证据,在多项高难度基准上实现稳健评判。本文系统研究了大语言模型在执行多标准监督与评判时的决策负载瓶颈及其安全风险。
提出对抗重构,仅改表述以拉高 AI 审稿分数
Published Jun 11
Evaluated models Claude Sonnet 4 , Claude Sonnet 4.5 , GPT-5-mini Abstract 揭示 AI 审稿系统易受纯表述重构操纵,建议在部署前解耦科学与写作评价。
提出 PaperGuard,用图文注入与梯度扰动抬高自动审稿分数并检测恶意指令
Published Jun 11
Evaluated models Claude-sonnet-4.5 , GPT-4o , GPT-4.1 and others11 Abstract PaperGuard 测到提示注入与白盒视觉抬分,并比较分块检索防御。PaperGuard 同时评测并检测多模态 AI 审稿中的评分操纵。
提出日志基底提示注入,操纵 SOC 分析员的分类、摘要与处置建议
Published May 23
Evaluated models gpt-4o-mini Abstract 在 gpt-4o-mini 上,人格劫持与上下文操纵可改写 SOC 分析输出,直接覆盖在分类上无效。
用简历中的不可见指令做间接提示注入,操纵招聘推荐
Published Feb 20
Evaluated models Qwen 3 30B A3B Instruct 2507 , Qwen 3 30B A3B Thinking 2507 Abstract 作者称推理在简单注入下会帮攻击圆谎,在不合逻辑的复杂注入下又会把欺骗漏出来。
测量 LLM 重排序器对文档内提示注入的脆弱性
Published Feb 18
Evaluated models Qwen3-0.6B , Qwen3-1.7B , Qwen3-8B and others16 Abstract 复现 DOH/DCH 对 LLM 重排序器的影响,并给出规模、位置、架构和领域上的边界。
提出隐蔽提示注入攻击 LLM 钓鱼检测器,并给出防御 InjectDefuser
Published Feb 5
Evaluated models OpenAI GPT-5 , xAI Grok 4 Fast (Non-Reasoning) , Meta Llama 4 Maverick and others4 提出查询无关黑盒攻击 DQ-A,用可迁移文末后缀降低文档检索排名
Published Jan 31
Evaluated models Qwen-1.5-7B-Instruct , SFR-Embedding-Mistral , E5-Mistral-7B-Instruct and others10 Abstract DQ-A 以无查询黑盒后缀压低 LLM 检索排名,并在 BRIGHT 上比较三种基线。DQ-A 是一种 query-agnostic、complete black-box 的文档隐藏攻击,对象是用最后隐层相似度匹配的 LLM 检索。