跳到正文
研踪PaperTrail

论文库

检索计算机领域论文,按研究方向、作者或关键词筛选。

部分采集异常
正在读取账号与研究记录…

全站更新概况

+0新增论文
+0其中 AI 相关
+0已有论文修订
+93新增 AI 导读

arXiv 最近完成扫描:2026-11-16 01:34 · 全站最近入库:2026-11-14 15:09各信源进度 →

统计口径与数据新鲜度

统计 2026-11-15 09:36 — 2026-11-16 09:36(北京时间) · 点击数字查看对应简报;AI 相关包含在新增论文内。

最近完成检查:2026-11-16 01:34最近发现新增版本:2026-11-14 15:09库内论文最新更新时间:2026-11-13 02:59

每分钟刷新已保存的数据。空检查不代表采集异常;新收录可能包含历史论文。AI 导读按新增版本计数。

23,348 已收录论文19,580 AI 相关6,729 有修订112 HF 社区发现

浏览论文库

时间范围跟随排序:版本更新、本站收录或首次发表。最新收录可能包含历史论文。部分信源异常,查看状态 →

找到 3 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-16 01:33(北京时间)
版本更新 · v2arXiv:2611.12067

XposedMIA: Exploiting Explainable Artificial Intelligence (XAI) for One-Shot Membership Inference Attacks

Omar Hamdache, Serdar Kara, Kıvanç Kuzey Dikici, Abdullah Çağlar Öksüz, Anisa Halimi, Erman Ayday, Sinem Sav

原文摘要

Machine learning (ML) services increasingly expose explainability outputs alongside model predictions. While such explanations are intended to improve transparency, they may also reveal information about the underlying training data. In this work, we show that the local surrogate geometry employed by Local Interpretable Model-agnostic Explanations (LIME) can be repurposed as a proxy for a sample's distance to the model's decision boundary. We extend this boundary-distance view to tabular counterfactual explanations and, for the first time, to visual counterfactual explanations for image classifiers. We present XposedMIA, a one-shot, explanation-based black-box membership inference attack (MIA) framework that converts the returned explanation into a proxy for the queried sample's distance to the model decision boundary. After a one-time calibration step on a small labeled calibration set of as few as 30 samples, our attack requires only a single query for each target sample. Unlike prior label-only attacks requiring iterative queries or shadow training, our method relies solely on the predicted label and its accompanying explanation. Our evaluation indicates that XposedMIA achieves high effectiveness, with success rates of up to 0.95 using only a single query per target, revealing that releasing explanations introduces an additional channel for membership leakage.

cs.LG

新论文 · v1arXiv:2611.12241

SNAPRAG: Cost-Efficient Knowledge Extraction Attack on RAG Systems via Adversarial Perturbations

Melih Coşğun, Mehmet Emre Gürsoy, Alptekin Küpçü, Sinem Sav

原文摘要

Retrieval-Augmented Generation (RAG) systems improve Large Language Model (LLM) responses by incorporating information from external knowledge bases. However, recent studies have shown that attackers can extract substantial portions of private knowledge bases from RAG systems through repeated querying. Yet, state-of-the-art extraction attacks either rely heavily on external LLMs to generate adversarial queries or require a large number of queries. In this paper, we propose SNAPRAG, a novel and cost-efficient attack for extracting RAG knowledge bases through adversarial perturbations. Our approach algorithmically generates adversarial queries without relying on external LLMs and allows for a strictly configurable query budget. In particular, SNAPRAG uses a discrete token-mutation algorithm to generate novel queries in a strict black-box setting. Our adaptive dissimilarity threshold calibration systematically forces queries into unexplored semantic regions of the target database, enabling extraction with fewer queries. Specifically, to extract half of the target knowledge base, our method requires up to 6.77$\times$ fewer queries than existing approaches. Furthermore, within a fixed query budget, our method recovers up to 96.9%, outperforming the closest approach by 7.7-19.9%. We further evaluate SNAPRAG against active defenses, where it remains the leading attack on four of five defended targets, achieving 50.7-96.3% extraction rate.

cs.CR

新论文 · v1arXiv:2611.04717

BrainwashedLLM: Jailbreaking via Automated Adversarial Chain of Thoughts Discovery

Masoud Poorghaffar Aghdam, Göktuğ Serdar Yıldırım, Sinem Sav

原文摘要

Large Reasoning Models (LRMs) achieve strong performance on complex tasks by generating explicit Chain-of-Thought (CoT) reasoning, but the safety implications of this intermediate reasoning surface remain underexplored. We demonstrate that the reasoning trace itself is a powerful attack vector. Safety alignment in large models is substantially less robust to malicious reasoning embedded in the context than to direct imperative prompts. We introduce BrainwashedLLM, a weak-to-strong adversarial framework that automates the discovery of reasoning-based jailbreaks. First, we use Group Relative Policy Optimization (GRPO), a Reinforcement Learning (RL) algorithm, with a multi-objective rule-based reward to fine-tune a small, white-box source model to generate adversarial reasoning traces. Crucially, the reward is computed entirely over the source's own reasoning trace. Second, we show that these traces transfer in a zero-shot manner. When traces are injected as input prefixes, they steer larger black-box LRMs to continue the adversarial reasoning and produce harmful outputs, without requiring prompt optimization, gradient access, or any form of output feedback from the target model. Across four state-of-the-art LRMs that span distinct architectures, BrainwashedLLM achieves ASR@5 between 0.74 and 0.98 across targets, within 16 points of query-based baselines, without querying the target during attack construction. To our knowledge, this is the first framework to use RL to directly optimize the CoT of LRMs as an attack surface without any target feedback, and the first to demonstrate zero-shot weak-to-strong transfer of reasoning traces. Our results suggest that exposed CoT is a vulnerability in LRMs and motivate architectural defenses that isolate the reasoning buffer. Warning: This paper contains potentially offensive and harmful text.

cs.CR