Extracting Forgotten Prompts from Targeted Unlearned Models

Recent unlearning methods (e.g. NPO, DPO, LUNAR) make use of refusal alignment to suppress forgotten data. However, it has been shown that refusal responses might leave traces of unlearning, and recent attacks have been able to successfully recover some of the unlearned knowledge. In this paper, we uncover a new vulnerability. Existing attacks typically assume that the forgotten prompts are already known to the adversary and focus on recovering their answers. However, we show that the forgotten prompts themselves can be extracted by using the retained data and black-box access to the model. Our attack, Targeted Active Search (TAS), first identifies the forgotten entities by constructing canonical templates and entity pool, and selectively querying the model using the most informative template-entity pair under a limited query budget. Once the entities are identified, TAS instantiates prompt templates with those entities to probe the unlearned model and reconstruct the forgotten prompts. Experiments across three unlearning methods with three datasets and three LLMs shows that TAS recovers the forgotten entity with $100\%$ accuracy and reconstructs up to $95\%$ of forgotten prompts, all while using up to $99.7\%$ fewer queries than naive probing.

Learning-Based Reconstruction Attacks on Coordinate-Obfuscated Point Clouds

Volumetric video based on point cloud representations enables immersive virtual and augmented reality applications but introduces significant challenges for efficient and secure content delivery. Prior work proposed a selective coordinate encryption framework for point clouds that encrypts only a subset of coordinates, reducing computational costs while visually degrading unauthorized content. However, it remains unclear whether the remaining unencrypted information is sufficient to enable content reconstruction. In this paper, we evaluate the robustness of selective coordinate encryption against machine learning-based reconstruction attacks. We consider an attacker with access to selectively encrypted point clouds attempting to recover encrypted coordinates without decryption by exploiting spatial and geometric correlations in the unencrypted data. We evaluate PointNet and Random Forest models under two encryption granularities: \texttt{X}, where all $X$ coordinates are encrypted, and \texttt{2X}, where every second $X$ coordinate is encrypted. Our results show that reconstructing fully encrypted $X$ coordinates remains challenging, whereas the \texttt{2X} scheme leaks sufficient information through neighboring coordinates to enable accurate reconstruction. These findings demonstrate that the security of selective coordinate encryption depends strongly on encryption granularity.

Privacy Leakage in Federated Learning: Gradient-Based Client Identity Inference and Defenses for Inertial Sensing in Vehicular Edge Networks

As vehicular networks move toward 5G/6G edge intelligence, federated learning (FL) is widely promoted as a privacy-preserving way for vehicles and infrastructure to train shared models without exposing raw sensor data. Yet the updates clients transmit still leak enough information to identify who sent them, which threatens the anonymity that safety-critical V2X applications assume and adds to existing concerns over adversarial ML, model poisoning, and backdoor attacks. We study server-side client identity inference from transmitted weight deltas using inertial (IMU) measurements, evaluated on the UCI Human Activity Recognition (HAR) benchmark as an accessible proxy for the IMU streams produced onboard connected vehicles. Across five attack classifiers and five non-IID partitions, an honest-but-curious server recovers client identity with near-perfect accuracy (approximately 1.000) from undefended updates, confirming a concrete identifiability risk. We then quantify the privacy-utility trade-off of a lightweight clip-then-noise defense by sweeping Gaussian noise (sigma in {0.00, 0.05, 0.10, 0.20, 0.50, 1.00}) at fixed clipping (C=1.0), and report formal (epsilon, delta)-DP budgets through Renyi accounting. A practical region (sigma in [0.1, 0.2]) drives attack accuracy to near-random while costing under 5% relative FL accuracy. Ensemble FL supplies complementary structural privacy with a 1/K anonymity-set bound and no noise penalty. Results are supported by cryptographic (SHA-256) train/evaluation gradient disjointness, three seeds, and a count-normalized attacker-advantage metric. We position HAR explicitly as a proxy and discuss what validation on true vehicular telemetry would require.

Hearing the Whispers: Black-Box Membership Inference Attacks on Finetuned TTS Models

Text-to-Speech (TTS) foundation models are increasingly fine-tuned on private datasets to synthesize highly personalized voices, introducing severe privacy risks by exposing both biometric identities and sensitive speech content. Existing black-box membership inference attacks (MIAs) follow a two-stage pipeline of query generation and representation engineering, both of which face unique challenges when adapted to TTS. For query generation, dual conditioning on synthesis text and reference speech creates a large and underexplored query design space with no established criterion for identifying an effective query. For representation engineering, the multi-level speech characteristics and temporal variability of speech make low-level representations and direct comparisons inadequate for capturing membership signals. To address these challenges, we present the first black-box MIA framework explicitly tailored to TTS models at both the speaker and record levels. For query generation, we characterize the feasible query space and establish two criteria, scorable extent and memorization elicitation, for evaluating five representative queries, identifying recitation as the strongest. For representation engineering, we obtain multi-level speech representations from embedding models and temporally align the generated and target audio for fine-grained comparison. Evaluations across three state-of-the-art TTS models (CosyVoice2, F5-TTS, and XTTS-v2) fine-tuned on two benchmark datasets (VCTK and British Dialect) reveal severe privacy leakage: speaker-level AUC remains above 0.80 and approaches 1.0 in the strongest settings, while record-level AUC ranges from 0.80 to 0.90 and remains effective even in challenging scenarios where both members and non-members are of the same speakers. We further identify speech characteristics associated with disproportionate vulnerability to memorization.

Membership Inference in Fine-tuned Diffusion Language Models via Token-level Memorization Asymmetry

Diffusion language models (DLMs) have recently emerged as an alternative modeling paradigm to autoregressive LMs, offering advantages such as parallel generation and bidirectional context modeling. Despite growing interest in their generative capabilities, the privacy risks of DLMs remain underexplored. We identify a phenomenon termed token-level memorization asymmetry through theoretical analysis of diffusion training dynamics. Building on this finding, we propose Q-Skew, a quantile-weighted skewness-based indicator for membership inference on finetuned DLMs. Experiments across multiple fine-tuning datasets and models show that our method outperforms existing baselines. Moreover, we show that Q-Skew can also facilitate other privacy violations, such as PII extraction. Our findings reveal a previously underexplored privacy attack surface and highlight the need for systematic privacy evaluation of DLMs.

Context Inference Attacks Without Jailbreaks

Agentic AI systems are increasingly deployed to process sensitive data at inference time, such as healthcare records or financial documents assembled into a hidden \emph{context} before the system answers. Prior work has studied privacy risks primarily through \emph{jailbreaking} attacks that induce models to directly disclose sensitive content, but has largely overlooked the agentic setting where the context is assembled by the agent's own tool calls. We show that the agents we evaluate remain vulnerable to hidden-context leakage despite the controls we test against them, namely an instruction not to disclose the context, logit suppression, and context dilution. For instance, a web-browsing agent answering benign user queries still carries exploitable signals about records silently loaded into its context. We introduce and formalize \emph{context-inference attacks} through a security game and evaluate three settings under decreasing attacker knowledge and increasingly indirect delivery of the context: a known context, an unknown context, and a context the agent retrieves through its own tool calls. We distinguish a grey-box setting, in which the target model is used to score observations, from black-box settings in which the attacker scores with a surrogate it controls. We further characterize how leakage varies with query budget, context size, and target-model size. A single attack carries through all three settings without modification, reaching $100\%$ ASR on small candidate sets and $63\%$ at $1024$ candidates against a known context, $78.9$ AUROC when the template and surrounding records are unknown, $92.5$ AUROC when a 14B surrogate scores a 32B target, and $81.8$ AUROC when the records arrive as an agent's retrieval returns, against chance rates of $1/|\mathcal{Z}|$ and $50$ respectively.

Curvature Cryptanalysis of Smooth Transformer Feed-Forward Networks

We show that smooth two-layer feed-forward networks (FFNs) expose an additional structural model extraction channel under a chosen-input raw-output oracle at the FFN branch; consider transformer FFN branches with GELU or SiLU activations under chosen-input raw-output access, without access to parameters, gradients, or internal activations; exploit a second-order leakage channel in which projected input Hessians form different mixtures of the same hidden symmetric rank-one factors induced by the FFN input weights. We formalize resulting Hessian collection as a partially symmetric decomposition to establish conditions for local identifiability and stability to exploit vector-output stencil reuse to reduce the structural query cost by a factor of 16. On independently trained CIFAR-10 vision transformers, only 16 projected Hessians, corresponding to 8193 black-box queries, recover the hidden FFN directions with average absolute cosine alignment above 0.94, with 95.1 % of GELU and 91.9 % of SiLU directions exceeding 0.90 alignment. Recovery remains high across independently trained models, repeated extraction runs, and all transformer blocks. The recovered structure supports functional extraction too. Keeping the recovered directions fixed and fitting only the remaining FFN parameters yields high-fidelity substitutes with more than 93 % top-1 agreement, while test accuracy remains within 0.90% and 0.62% of the GELU and SiLU targets. Output rounding and Gaussian noise substantially reduce recovery under a fixed attack configuration, but adapting the finite-difference step restores average alignment to 0.9603 and 0.9398. This is an end-to-end path from black-box second-order observations to hidden FFN-structure recovery and functional replacement. Under the stated oracle model, smooth FFN curvature exposes internal parameter geometry that behavioral fidelity alone cannot reveal.

DIME: Query-Efficient Framework for Membership Inference on Diffusion Models

Membership inference attacks expose whether individual records were used to train a model, yet existing attacks on diffusion models are largely heuristic and can require substantial query budgets. We introduce DIME (Denoiser Ideal Membership Error), a theoretically grounded and query-efficient framework for membership inference on diffusion models. Our starting point is an exact characterization of the optimal diffusion denoiser for a finite training set, which reveals that membership leakage is governed by the denoiser's implicit reconstruction error. This error decomposes into two complementary signals: a bias term, capturing reconstruction accuracy, and a previously unexplored local crowding term, capturing the geometry of nearby training examples. Both admit efficient estimators using only model queries, yielding a practical attack with as few as two queries. Across CIFAR-10/100, STL10-U, CelebA, and ImageNet, DIME consistently outperforms prior attacks at comparable or substantially lower query cost, improving TPR at 1% FPR by up to $3\times$; remarkably, its two-query variant can outperform existing 30-query baselines. Finally, we suggest, discuss, and evaluate specific defenses to counteract such powerful membership tests.

Denoising-Aware Inversion: Revealing Privacy Risks in Noise-Protected Text Embeddings

Dense text embeddings are widely used in data mining, retrieval, and downstream machine learning systems due to their compact and semantically rich representations, but recent embedding inversion attacks have shown that they can expose substantial information about the original text, leading to serious privacy leakage risks. A common defense is to release perturbed embeddings by adding Gaussian noise, which is simple yet effective against standard inversion attacks and does not significantly degrade embedding utility for downstream tasks. However, it remains unclear whether such noise-protected embeddings are sufficiently safe against adaptive attackers that explicitly account for the perturbation process. In this paper, we study text embedding inversion in a noise-protected setting, where the attacker can observe only noisy embeddings and has no access to clean embedding targets. We first analyze why existing generative inversion methods fail under this setting and identify a "Double Noise Trap", which fundamentally prevents standard generative inversion models from achieving high-quality reconstruction. To address this challenge, we propose DAEI, a denoising-aware embedding inversion pipeline that combines a residual denoising autoencoder with generative text inversion where the denoiser is trained in an unsupervised manner using Stein's unbiased risk estimate to enable denoising from noisy observations alone. Extensive experiments show that DAEI achieves approximately 154\% relative improvement in BLEU over the existing generative inversion baseline, while also improving token-level F1 and ROUGE-L by 32--60\%. The promising inversion performance of DAEI challenges the prevailing assumption that simple Gaussian perturbation is sufficient to prevent sensitive information leakage from embedding representations.

Steering the Flow: Inverting Face Recognition Models via Gradient-Guided Flow Matching

Model Inversion Attacks (MIAs) aim to reconstruct representative training samples of target identities from face recognition models, exposing critical security vulnerabilities. Existing methods typically rely on indirect guidance or highly stochastic guidance, making it difficult to stably optimize generation trajectories toward target facial images. In this paper, we propose Steering Flow Model Inversion (SFMI), a novel two-stage white-box model inversion method that reformulates inversion as a trajectory-steering task. Specifically, Step I, Learning a Generic Flow Matching Prior, pre-trains a generic unconditional Flow Matching model to encode the manifold of human faces as a robust prior. Step II, Attacking with Progressive Guidance Scheduler (PGS), injects time-dependent target-specific gradients during sampling. By backpropagating through the target model to obtain gradients from intermediate generated states, PGS progressively injects adaptive guidance signals into the vector field. This process effectively steers the current generative flow from random noise toward the high-density regions of the target class. Under an identity-disjoint cross-evaluation setting using the CelebA dataset, SFMI achieves an ACC of 0.9248, an FID of 22.61, and an LPIPS of 0.3874 on the ArcFace target. Extensive experiments on multiple target models demonstrate that SFMI achieves competitive state-of-the-art performance in attack success and visual fidelity under the evaluated white-box protocol.

AccretionLink: On-Device Auditing of Exposure-Control Attacks on Attribute Inference

Exposure control lets an adversary rank authentic public posts to strengthen private-attribute inference without altering content. AccretionLink defines confidentiality and integrity games for this attack, models bounded selection odds through partial identification, and constructs dependence-aware time-uniform e-processes. On 52 held-out synthetic profiles, odds-four selection reduced aggregate negative log likelihood at every horizon. At eight posts the advantage was 0.01595 nats (95% CI [0.00890, 0.02336]), three of four target effects survived Holm adjustment, and label-blind model-guided selection caused 6/109 high-confidence false reversals. On 142 PAN15 test profiles, exploratory selection produced a 0.01227-nat advantage but no reversal. A separate TF-IDF selector retained a 0.01470-nat advantage against the unchanged G5 target, while matched identity shuffling did not reproduce it. Pixel 10 encoded all 1,622 held-out posts once with a fallback-free Tensor G5 graph. A P-256 checkpoint authenticated the selected-replay, actual-model, native-report, and operation digests; local KeyInfo identified the signing key as StrongBox-backed.

Stealing Reasoning Traces from Proprietary LLM APIs

Leading large language model providers now conceal their models' step-by-step reasoning, or chain-of-thought, to protect intellectual property and limit information leakage. Rather than storing these traces server-side, providers return them to the client as blocks of encrypted text, which the client passes back with each subsequent request. Building on prior research, we identify an architectural vulnerability: these encrypted blocks are fully compatible and interchangeable across different sessions, users, and models within a provider's ecosystem. We exploit this compatibility to develop a scalable decryption jailbreak. By injecting an encrypted reasoning trace from a given model into a weaker, and less safeguarded model from the same provider, we force it to decode and output the trace verbatim in plaintext, without ever jailbreaking the more capable model directly. This vulnerability enables four distinct attack vectors. First, it circumvents anti-distillation mechanisms, allowing adversaries to extract a proprietary model's reasoning, as we demonstrate across Anthropic, OpenAI, and Google. Second, it allows for large-scale private data extraction. Developers frequently share session logs publicly, unaware of contents of the encrypted blocks. By decoding 315,320 reasoning blocks scraped from public repositories, we recovered 367 Personally Identifiable Information (PII) artifacts and 182 credentials. Third, it inadvertently reveals hazardous information hidden within the reasoning process, even in cases where the model's final, visible output safely rejects a malicious request. Fourth, attackers can leverage this flaw to execute invisible prompt injections, embedding malicious payloads entirely within encrypted blocks to poison public agentic rollouts. Following responsible disclosure, we propose concrete cryptographic and system-level mitigations to secure client-side reasoning.

DeepInvert: Semi-Supervised Embedding Inversion Against Obfuscated Language Models

Cloud-based language model services routinely process prompts containing sensitive information. Obfuscation-based defenses---including ObfusLM, SentinelLMs, TextObfuscator, and DPNR---mitigate this risk by transforming prompt representations before transmission, offering a lightweight alternative to cryptographic solutions. We show these defenses provide far less protection than previously believed. We present DeepInvert, a semi-supervised embedding inversion attack that recovers original tokens from obfuscated representations with higher accuracy than prior methods. The key insight is that unlabeled obfuscated embeddings retain exploitable semantic structure despite perturbation. DeepInvert combines supervised training on labeled shadow data with a novel unsupervised consistency objective over unlabeled target embeddings, alternating between the two via a mixed training pipeline. Defense-aware adaptations further extend the attack to diverse obfuscation mechanisms across encoder-based and autoregressive architectures. Experiments on nine defenses, five tasks, and four model architectures show that DeepInvert outperforms prior attacks on most defenses. Against ObfusLM, DeepInvert achieves 73.5\% top-1 token recovery versus 26.2\% for the previous best. Our results reveal a task-dependent tension: obfuscation schemes preserving enough signal for utility also retain sufficient structure for inversion, while schemes resisting inversion collapse utility. On simpler classification tasks, some DP-based defenses can maintain both. We call for a re-evaluation of this defense class.

MineGrad: Gradient Inversion Attacks on LoRA Fine-Tuning

Parameter-efficient fine-tuning (PEFT), such as low-rank adaptation (LoRA), has recently been adopted in federated learning to reduce communication and computation costs. In this setup, users download a pretrained model from the server prior to fine-tuning, and then fine-tune lightweight LoRA modules locally while keeping the pretrained model frozen, sharing only the gradients of the fine-tuning parameters with the server. Despite its growing popularity, robustness of federated fine-tuning against an adversarial server remains underexplored, where the server maliciously tampers with the training protocol to breach the privacy of users' data. In this work, we investigate gradient inversion attacks on LoRA fine-tuning. We propose an analytical attack that enables a malicious server to recover private user data by leveraging a poisoned pretrained model and fine-tuning parameters. Our design embeds fine-tuning data within the shared gradients, to allow the server to analytically reconstruct user data. Unlike prior works, our attack is applicable to both language and vision tasks, does not rely on computationally expensive (adversarial) pretraining with public datasets or require the number of training tokens to be less than the rank of LoRA modules. Experimental results on both language and vision tasks demonstrate high-fidelity data recovery across multiple baselines, revealing several critical vulnerabilities.

Do Speech Tokens Leak Voiceprints? Speaker Inversion Attacks Against End-to-End Speech Language Models

End-to-end speech language models increasingly represent user speech with speech tokens rather than relying exclusively on cascaded ASR--LLM--TTS pipelines. Although these tokens support expressive and low-latency spoken interaction, they may also preserve sensitive speaker characteristics. We investigate whether exposed speech tokens leak voiceprints and formulate this risk as a speaker inversion attack. We introduce Audio BERT (AuB), a trainable model that constructs token embeddings from discrete codebooks and aggregates them into speaker-sensitive representations, and propose SpInv, a two-stage inversion method built on AuB to recover embeddings in the space of an attacker-specified speaker encoder. We evaluate Moshi, Higgs3, Kimi-Audio, and Qwen3-Omni using speaker-disjoint protocols on the VoxCeleb dataset. Extensive experiments show that, with only three seconds of frontend output, SpInv achieves cosine similarities above 0.70 in the attacker-specified speaker-encoder space.

FedCVESA: Taking Away Training Data in Federated Learning via Correlation Value Encoding and Segmented Aggregation

Federated learning (FL) avoids explicit data exposure by keeping raw data on local clients, yet privacy risks remain in the training process and the learned model itself. Recently, centralized Taking Away Training Data (TATD) attacks have shown that malicious training could abuse the memorization capacity of deep models to store and later recover training data. However, this memorization-based threat has not been systematically studied under FL environments, where multi-client averaging could overwrite encoded training data. In this paper, we study a white-box TATD attack in which a malicious server selects n target clients from K participating clients and actively writes private training data into the global model during federated training. We propose FedCVESA, a federated variant of Correlation Value Encoding Attack (CVEA), by adding a Pearson-correlation regularizer to the loss function of target clients, so that private training data are gradually encoded into selected model parameters, referred to as carrier parameters. To reduce the overwriting of carrier parameters during server aggregation, we further propose segmented aggregation over dispersed carrier parameters, preserving selected carrier parameters while keeping standard averaging on the remaining parameters. Experiments on MNIST, Fashion-MNIST, and CIFAR-10 under Dirichlet non-IID partitions show that the proposed method can steal semantically meaningful private training images from the trained model while maintaining acceptable main-task utility in a controlled proof-of-concept setting. These results demonstrate that FL can become a parameter-level memorization channel for active TATD attack under the studied white-box malicious-server setting.

POPS: Recovering Unlearned Multi-Modality Knowledge in MLLMs with Prompt-Optimized Parameter Shaking

Multimodal Large Language Models (MLLMs) have demonstrated impressive performance on cross-modal tasks by jointly training on large-scale textual and visual data, where privacy-sensitive examples could be unintentionally encoded, raising concerns about privacy or copyright violation. To this end, Multi-modality Machine Unlearning (MMU) was proposed as a mitigation that can effectively force MLLMs to forget private information. However, the robustness of such unlearning methods is not fully exploited when the model is published and accessible to malicious users. In this paper, we propose a novel adversarial strategy, namely Prompt-Optimized Parameter Shaking (POPS), aiming to recover the supposedly unlearned multi-modality knowledge from the MLLMs. Our method elicits the victim MLLMs to generate potential private examples via prompt-suffix optimization, and then exploits these synthesized outputs to fine-tune the models so they disclose the true private information. The experiments on the different MMU benchmarks reveal substantial weaknesses in the existing MMU algorithms. Our POPS can even achieve a near-complete recovery of supposedly erased sensitive information on the unlearned MLLMs, exposing fundamental vulnerabilities that challenge the foundational robustness of representative MMU-based privacy protections.

Black-Box Inference of LLM Architectural Properties with Restrictive API Access

In practice, most commercial LLM providers do not publicly release details of underlying LLM architectures. However, prior work has shown that given limited API access to an LLM (namely, top-$k$ logits and/or a logit bias function), one can recover certain architectural details of an LLM, such as the hidden dimension of the feed-forward network. Perhaps in response to these results, most commercial LLM providers have restricted their APIs to expose only the single logit for each decoded token, and they no longer give users the ability to bias logits. We show that even under current restrictive APIs, several architectural parameters are still recoverable. We present NightVision, an attack that uses restrictive black-box API access to estimate the hidden dimension, depth, and parameter count of an LLM. Algorithmically, NightVision relies on a novel common set prompting technique in which multiple prompts expose log probabilities for the same set of output tokens; a spectral analysis of these results is used to infer hidden dimension. NightVision additionally uses end-to-end time to first token (TTFT) measurements and the estimated hidden dimension to estimate depth and parameter count. We empirically evaluate NightVision on 32 open-source LLMs, recovering hidden dimension to within 23% average relative error across all models (9% on MoE models), and depth and parameter count to within 53% for models exceeding three billion parameters. We run extensive ablations to demonstrate how these accuracies scale with token budget and model properties. Overall, our results suggest that current LLM APIs are not sufficiently restricted to fully obfuscate the architectural details of their underlying models.

Recovering Input Text from Hidden States: Study of Gradient-Based Inversion of Decoder-Only Language Models

This work studies the hidden-state inversion problem: recovering the original input token sequence of a decoder-only language model from its last-layer hidden states. Rather than treating inversion as a one-shot reconstruction, we study it as a continuous embedding-space optimisation in which a soft proxy is driven towards the leaked target without any hard-token projection during the search, and a token is committed only once, at the end of the inner loop. This design choice has two consequences which are the main focus of this paper. First, keeping the optimisation entirely in continuous space exposes a rich set of internal signals: rank trajectories of the ground-truth token, per-position loss curves, and a discrete loss measured at commit time. Second, the discrete loss allows assessing the correctness of recovery via cumulative discrete loss. We further analyse which tokens break the reconstructions and find a sharp categorical asymmetry: space-prefixed, high-frequency function words in dense regions of the embedding matrix dominate the failures, while content-bearing tokens are recovered almost perfectly. On 10-token C4 prompts the exact-match rate rises from 66.9% to 97.5% (mean similarity 0.994) as the candidate window is widened, confirming that most errors are recoverable near-misses rather than genuine ambiguities. A comparison with the released SIPIT reference situates these findings: per-step hard projection is faster, but the continuous formulation is what makes the optimisation observable and its failures detectable. The results show that last-layer hidden states of GPT-2 are as sensitive as the original text.

Seeing Through the Weights: Privacy Leakage in Scene Coordinate Regression

Scene Coordinate Regression (SCR) methods are increasingly adopted for visual localization. In these approaches, the scene is implicitly encoded within a neural network that regresses a 3D world coordinate for each image pixel. Because the scene is represented only through the network parameters and not stored explicitly as images or maps, such methods are often assumed to be privacy-preserving. In this work, we show that this assumption is incorrect in practice. Specifically, we introduce a query-based attack that reconstructs the 3D geometry of the training environment from an SCR model under different levels of model access. To do so, we repeatedly query the model with batches of proxy images unrelated to the target scene to obtain dense pixel-wise 3D coordinates. Reliable points are identified through their stability under small input perturbations and can be further refined in a white-box setting. These stable points are accumulated across independent query batches to recover the scene geometry. From the recovered 3D representation, we also invert the network features to synthesize images from arbitrary viewpoints, revealing additional appearance information. Experiments on indoor and outdoor datasets demonstrate that substantial portions of training environments can be reconstructed with high geometric fidelity. Beyond geometry, we also recover an approximate color appearance, which exposes recognizable layout and potentially sensitive scene elements. This directly contradicts claims in the literature that SCR representations are privacy-preserving by design, and reveals a real risk when such systems are deployed in private or security-critical spaces. The project page is available at https://jaeminch0.github.io/seeing-through-the-weights-privacy-leakage-in-scene-coordinate-regression.

Rethinking Generative Reconstruction Attacks against Graph Neural Network Models

The application of graph data in numerous disciplines raises the need for gathering and analyzing huge volumes of data, some of which is private and sensitive. The non-Euclidean nature of the graph data makes the analysis computationally challenging, leading to the use of Graph Neural Networks (GNNs) in the age of AI. GNNs may inadvertently leak sensitive data they are trained on, which raises serious data security issues, including the model inversion attack. In this study, we analyze GNNs' vulnerabilities by introducing two novel graph inversion (i.e., reconstruction) attacks: graph-label conditioned (GLC) attack and embedding-label conditioned (ELC) attack, utilizing targetmodel predictions and their intermediate representations, respectively. We perform a comprehensive analysis of our introduced privacy attacks and compare them with existing baselines across three benchmark graph datasets (i.e., NCI1, PROTEINS, and AIDS) and four graph distributional/structural metrics (i.e., FGD, EGD, MMD, and GKS). Our work demonstrates that an adversary can use the generator-discriminator technique to reconstruct high-quality graphs in real-world black-box attack scenarios against GNNs. Additionally, we present a variant of our attacks (Ours--) with 50% reduced queries, achieving good or comparable reconstruction attack performance. In addition, we show that GNNs are highly vulnerable to privacy attacks, varying Laplacian noise-scales.

Agentic AI-Powered Re-Identification: An Emerging, Scalable Threat to Mobility Microdata Privacy

The widespread collection of fine-grained location data by commercial data brokers creates a re-identification risk that is not widely recognised by the public. While prior research has established that mobility traces are highly unique and that individuals can, in principle, be identified from a handful of spatio-temporal points, such attacks have historically required significant manual effort from skilled analysts, limiting their practical scale. In this feasibility study, we demonstrate in a real world setting that agentic AI fundamentally changes this threat model. We present an end-to-end pipeline in which large language model agents autonomously search the open web, cross-reference public records and social media, and resolve raw coordinate sequences to candidate identities - without human intervention. We evaluate the pipeline on a spatio-temporal dataset containing simulated location points anchored at and around true home and work addresses, focusing on a high-risk disclosure scenario. Our results demonstrate that, from spatio-temporal data and public sources alone, our agentic AI successfully re-identified 18 of the 25 re-identifiable individuals (72%) and 18 of 43 cases overall (41.9%). We discuss implications for Statistical Disclosure Control (SDC) practice and outline the near-future escalation that data custodians and regulators must anticipate. De facto anonymity - an implicit foundation of SDC practice - is shifting. Agentic AI strengthens the case that re-identification is reasonably likely by any means under the GDPR Recital-26 standard, at costs of minutes-and-dollars per target.

Privacy Vulnerabilities of Attention Layers in Tabular Foundation Models and Protection of High-Risk Queries

Tabular foundation models are commonly assumed to present limited privacy concerns as they are often pre-trained on large collections of synthetic data. However, these models leverage in-context learning, where sensitive records may be provided directly at inference time as labelled context examples. In this paper, we demonstrate that predictions generated via the attention mechanism leak sufficient information to enable effective Membership Inference Attacks (MIAs). To highlight this vulnerability, we propose AMIA (Attention-based Membership Inference Attack), a shadow-model-free attack that exploits the concentration of transformer attention patterns. Our results show that attention mechanisms reveal strong membership signals, which exceed classical confidence-based attacks, achieving an average gain of 7.7\%, specially in low false-positive regimes. To mitigate this risk, we introduce an inference-time defence inspired by $k$-anonymity principles. This approach reduces the uniqueness of context-key representations without introducing random noise or retraining the model. By targeting only high-risk queries identified through AMIA scores, the defence substantially reduces membership leakage of this attack by an average of 50\% and 25\% against confidence-based attacks, while preserving predictive utility with only 3.9\% performance degradation. Beyond showing that context examples are vulnerable, we further demonstrate that fine-tuning introduces an additional source of privacy risk. In particular, samples whose prediction confidence increases after fine-tuning become more susceptible to MIAs, indicating that fine-tuning can amplify memorisation and expose sensitive training information through confidence shifts.

TIGER: Inverting Transformer Gradients via Embedding-Subspace Distance Optimization

Federated learning allows multiple clients to jointly train a shared model by sending gradient updates to a central server while keeping raw inputs local. However, prior gradient inversion attacks show that these updates can reveal enough information to reconstruct client inputs. Existing attacks on transformers either optimize dummy inputs to match the true client updates, which is costly and unstable for modern models, or exploit the low rank of attention gradients to identify a subspace containing the true layer embeddings, followed by a discrete membership test for candidate tokens. However, this token test is brittle under numerical noise, i.e., from quantization or Differential Privacy (DP), and scales poorly for encoder models with non-causal attention. We introduce TIGER, a continuous gradient inversion attack that turns this subspace signal into a differentiable objective. Instead of searching over tokens or matching full gradients, TIGER directly optimizes token embeddings to minimize their distance to the subspace. Our experiments demonstrate that on encoder-only models, TIGER substantially improves both reconstruction quality and runtime over existing attacks, while on decoder models, TIGER is more robust than prior subspace-based attacks, enabling the first successful reconstructions in DP-defended federated learning settings.

ResAware: Cross-Environment Website Fingerprinting via Resource-Privileged Distillation

While Website Fingerprinting (WF) attacks achieve high accuracy in controlled laboratory settings, they often degrade substantially in real-world environments due to spatio-temporal drift, browser heterogeneity, proxy obfuscation and etc. This limitation stems from their sole reliance on low-level traffic features that are noisy and highly sensitive to environmental perturbations. To address this problem, we propose \textbf{ResAware}, a cross-environment resource-aware distillation framework under a \textit{training-rich/inference-poor} asymmetric setting. Specifically, ResAware trains a teacher model on resource-level features, and then distills the resulting privileged knowledge into a student model through heterogeneous knowledge distillation. At deployment time, the student model performs inference using only encrypted traffic, incurring zero additional cost. We evaluate ResAware on a large-scale dataset collected over five months from six globally distributed vantage points, comprising more than $160{,}000$ paired samples. The results show that ResAware significantly enhances the cross-environment robustness of diverse WF baselines. Under a 150-day temporal drift, for example, ResAware improves the F1-score of Var-CNN from $72.77\%$ to $81.49\%$ and the open-world $TPR@1\%FPR$ from $22.40\%$ to $27.20\%$. Our results demonstrate that resource-level supervision improves WF robustness without expanding online observation capabilities.

Loss Landscape Poisoning: Targeted Extraction of Unseen Training Data from LLMs

Large Language Models are increasingly trained on proprietary or sensitive data, from private healthcare and financial records to user conversations containing secrets. Ensuring the privacy of such data against extraction attacks has become a central concern. In this paper, we ask whether an attacker who can poison a portion of the training data can facilitate the leakage of a separate target record they have no access to. We answer in the affirmative and show that such leakage can be induced by a poisoning mechanism that reshapes the model's local loss landscape around the target completion. Our key insight is that poisoning to create a sharp loss minimum at the target, surrounded by elevated loss on nearby alternatives, forces the model to memorize the target as the unique low-loss solution in its neighborhood. The attack requires no architectural changes, and generalizes across centralized and federated learning settings. We demonstrate that the attack amplifies privacy leakage across language (up to 100% successful extraction), and vision-language models (up 90% successful extraction). We show that the attack is thwarted when the model is trained to be differentially private. However, we introduce a new attack that directly probes the loss landscape bypassing even differential privacy defenses.

Attribute Inference from Interactive Targeted Ads

Targeted advertising systems can pair audiences selected by advertisers with ad units that expose visible user actions. When an interaction remains linked to the campaign that elicited it, the advertiser may receive an observation tied to a user rather than only an aggregate report. We model that channel as a noisy oracle for attribute inference. The model separates targeting predicates, exposure, interaction, and disclosure. These boundaries capture the gap between eligibility and delivery, and the gap between interaction and advertiser visibility. We build a reproducible benchmark using synthetic populations calibrated with public data, each with known sensitive labels. A generated campaign semantics layer provides topic variants and response priors. The simulator generates the ground truth, event traces, disclosed observations, and metrics. The evaluation compares Bayesian, supervised, positive and unlabeled, and adaptive attacks under common campaign and disclosure definitions. The final evaluation uses four topic variants, seven simulator seeds, and two interaction settings. Repeated campaigns with identity exposure produce measurable but bounded inference signal. At $160$ campaigns, Bayesian and supervised attacks reach about $0.64$ AUC in the main setting and about $0.65$ AUC in the higher interaction setting. Disclosure policy is the strongest control. Aggregate reporting removes the evaluated oracle input tied to users. Type filtering and randomized disclosure reduce the released signal. The result is a model, artifact, and defense evaluation method for privacy in interactive targeted advertising. The code is available at https://github.com/P-HOW/Interactive-Ad-Oracle.

From Prompts to Responses: Dual-Sided Data Leakage and Defense in Split Large Language Models

Large language models (LLMs) are increasingly deployed in privacy-sensitive domains, where users must balance the risk of data exposure through external APIs against the high computational cost of local deployment. Split learning has therefore emerged as a promising paradigm for LLM fine-tuning and inference under limited local resources. However, it introduces new privacy risks. Prior work primarily studies leakage of private input prompts, typically via inversion attacks on intermediate representations, while the potential for sensitive information leakage through generative response outputs remains largely unexplored. In this work, we unveil novel vulnerabilities of Split-LLM by presenting Patched Model Inversion with Dual-Sided Initialization (PIDI), a two-stage attack that simultaneously targets both private input prompts and output responses in Split-LLM settings. It combines dual-sided initialization with a patched inversion strategy to tackle long sequences, substantially outperforming prior inversion methods. To counter threats from both sides, we further propose the Adapter-based DualGuard with Mutual Information Defense (ADMI), which integrates an adapter-based local warmup strategy and mutual information regularization to provide a strong empirical privacy protection with minimal impact on task performance. Extensive experiments across diverse tasks and models demonstrate that ADMI effectively defends against PIDI and other state-of-the-art inversion attacks. Our code is publicly available at https://github.com/FLAIR-THU/VFLAIR-LLM.

The Vision Encoder as a Privacy Boundary: Visual-Token Side Channels in Encoder-Free Vision-Language Models

A vision encoder compresses image pixels into semantic embeddings, implicitly acting as a privacy boundary by preserving semantic content while attenuating pixel-local detail required for exact text recovery. Encoder-free vision-language models (VLMs) remove this boundary by routing image patches directly into the language-model token stream, thereby exposing an architectural privacy attack surface: intermediate visual tokens become a pre-output side channel. Under a token-access adversary, decoders invert visual-token streams from two encoder-free VLMs, Gemma4 and Fuyu, recovering recognizable image structure and readable held-out access codes, whereas matched encoder-based controls localize target regions but recover no exact strings. Within-model ablations show that the operative factor is spatial sampling fidelity of the visual-token grid, especially character-direction sampling density, rather than token or value count. The leakage is not limited to exported tokens: Gemma4 layer-0 key-value cache tensors are directly invertible, placing the side channel within KV caches commonly persisted by production serving stacks for decoding efficiency. The attack survives clutter, realistic document degradation, and zero-shot transfer to public document images, and it resists value-level defenses such as additive noise and quantization. Effective mitigation must therefore reduce spatial sampling, making removal of the vision encoder a first-class privacy decision in VLM deployment.

From Beats to Breaches:How Offensive AI Infers Sensitive User Information from Playlists

The pervasive integration of AI has enabled Offensive AI: the exploitation of AI for malicious ends across the cyber-kill chain. A critical manifestation is the user attribute inference attack, where AI infers sensitive Personally Identifiable Information (PII) from innocuous public data. We explore how music streaming ecosystems, where users routinely release public playlists, can be exploited for Offensive AI. To quantify this threat, we developed musicPIIrate. This novel tool leverages deep learning architectures that utilize both standalone data representations and the structural information embedded in a user's playlist collection. Our design explores set-based approaches (e.g., Deep Sets) and methodologies modeling relationships between playlists (e.g., Graph Neural Networks), which we also combine to leverage both perspectives. Our approach addresses feature extraction from unordered, variable-length set data, enabling accurate PII prediction. Empirical evaluation demonstrates that musicPIIrate achieves state-of-the-art inference accuracy. The tool successfully infers a wide array of attributes, including: Demographics (Age, Country, Gender), Habits (Alcohol, Smoke, Sport), and Personality Traits (OCEAN scores). musicPIIrate outperforms existing methods, beating baselines in 9 out of 15 attribute inference tasks. To counter this vulnerability, we propose JamShield, a lightweight defensive framework. JamShield strategically injects dummy playlists into an account to dilute the PII-carrying signal. Our analysis indicates that JamShield represents a promising defense, lowering inference F1-scores by an average of 10%. This work provides an initial Offensive-AI benchmark for playlist-based PII inference using architectures that leverage set- and graph-structured data and introduces a defense showing encouraging mitigation effects.