arXiv subjects

Sanghyun Hong

Publications and source records attributed to Sanghyun Hong.

At least 19 recordsLinked to original sources

arXiv2026

Do Reasoning Representations Help Humans Evaluate LLM Outputs?

Reasoning representations are increasingly used as explanations for large language model outputs. Yet they are typically evaluated with model-centric criteria, such as answer accuracy and faithfulness, leaving it unclear whether they help people evaluate model responses. In this work, we study reasoning representations as human-facing interfaces rather than proxies for model reasoning ability. We conduct a controlled human study of six reasoning formats across tasks of varying complexity, supported by a web-based framework that randomizes task domains, problem instances, and representation order. The study collects fine-grained judgments of structural understanding, error detection and localization, and trust calibration. Our study shows a mismatch between perceived preference and support for human evaluation. Participants prefer planning- and decomposition-based representations, but simpler chain-of-thought traces better support verification, trust, and interpretability. Preferred representations also introduce calibration risks, with more false alarms on correct traces and high trust despite low willingness to verify.

cs.LG↗

arXiv2026

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.

cs.CL↗

arXiv2026

On the Resilience of Text-to-Video Diffusion Models to Hardware Faults

We present the first systematic study of the resilience of text-to-video (T2V) diffusion models under random hardware-level faults. While T2V models are widely used for automated video generation due to their ability to produce high-quality, temporally coherent, and realistic videos, their iterative denoising process and spatiotemporal dependencies introduce unique failure modes. We perform an extensive fault-injection study covering both computational and memory faults across three T2V models and a representative benchmark. Our results show that (1) a single fault can degrade overall performance by up to 3.7\%, with semantic correctness more affected than perceptual quality; (2) memory faults are more damaging than computational faults, high-order exponent bits are particularly vulnerable, and the widely-used bfloat16 is more susceptible than alternative formats; and (3) 7-28\% of faults cause visible artifacts, including semantic changes such as added objects, suggesting that single faults are sufficient to alter output semantics. Our findings reveal reliability risks in deployed T2V systems and motivate further research on improving fault resilience. Code: \href{https://github.com/ztcoalson/T2V-Resilience}{https://github.com/ztcoalson/T2V-Resilience}.

cs.LG↗

arXiv2026

When Can You Poison Rewards? A Tight Characterization of Reward Poisoning in Linear MDPs

We study reward poisoning attacks in reinforcement learning (RL), where an adversary manipulates rewards within constrained budgets to force the target RL agent to adopt a policy that aligns with the attacker's objectives. Prior works on reward poisoning mainly focused on sufficient conditions to design a successful attacker, while only a few studies discussed the infeasibility of targeted attacks. This paper provides the first precise necessity and sufficiency characterization of the attackability of a linear MDP under reward poisoning attacks. Our characterization draws a bright line between the vulnerable RL instances, and the intrinsically robust ones which cannot be attacked without large costs even running vanilla non-robust RL algorithms. Our theory extends beyond linear MDPs -- by approximating deep RL environments as linear MDPs, we show that our theoretical framework effectively distinguishes the attackability and efficiently attacks the vulnerable ones, demonstrating both the theoretical and practical significance of our characterization.

cs.LG↗

arXiv2026

Fail-Closed Alignment for Large Language Models

We identify a structural weakness in current large language model (LLM) alignment: modern refusal mechanisms are fail-open. While existing approaches encode refusal behaviors across multiple latent features, suppressing a single dominant feature$-$via prompt-based jailbreaks$-$can cause alignment to collapse, leading to unsafe generation. Motivated by this, we propose fail-closed alignment as a design principle for robust LLM safety: refusal mechanisms should remain effective even under partial failures via redundant, independent causal pathways. We present a concrete instantiation of this principle: a progressive alignment framework that iteratively identifies and ablates previously learned refusal directions, forcing the model to reconstruct safety along new, independent subspaces. Across four jailbreak attacks, we achieve the strongest overall robustness while mitigating over-refusal and preserving generation quality, with small computational overhead. Our mechanistic analyses confirm that models trained with our method encode multiple, causally independent refusal directions that prompt-based jailbreaks cannot suppress simultaneously, providing empirical support for fail-closed alignment as a principled foundation for robust LLM safety.

cs.LG↗

arXiv2026

Discovering Universal Activation Directions for PII Leakage in Language Models

Modern language models exhibit rich internal structure, yet little is known about how privacy-sensitive behaviors, such as personally identifiable information (PII) leakage, are represented and modulated within their hidden states. We present UniLeak, a mechanistic-interpretability framework that identifies universal activation directions: latent directions in a model's residual stream whose linear addition at inference time consistently increases the likelihood of generating PII across prompts. These model-specific directions generalize across contexts and amplify PII generation probability, with minimal impact on generation quality. UniLeak recovers such directions without access to training data or groundtruth PII, relying only on self-generated text. Across multiple models and datasets, steering along these universal directions substantially increases PII leakage compared to existing prompt-based extraction methods. Our results offer a new perspective on PII leakage: the superposition of a latent signal in the model's representations, enabling both risk amplification and mitigation.

cs.LG↗

arXiv2026

Asking Forever: Universal Activations Behind Turn Amplification in Conversational LLMs

Multi-turn interaction length is a dominant factor in the operational costs of conversational LLMs. In this work, we present a new failure mode in conversational LLMs: turn amplification, in which a model consistently prolongs multi-turn interactions without completing the underlying task. We show that an adversary can systematically exploit clarification-seeking behavior$-$commonly encouraged in multi-turn conversation settings$-$to scalably prolong interactions. Moving beyond prompt-level behaviors, we take a mechanistic perspective and identify a query-independent, universal activation subspace associated with clarification-seeking responses. Unlike prior cost-amplification attacks that rely on per-turn prompt optimization, our attack arises from conversational dynamics and persists across prompts and tasks. We show that this mechanism provides a scalable pathway to induce turn amplification: both supply-chain attacks via fine-tuning and runtime attacks through low-level parameter corruptions consistently shift models toward abstract, clarification-seeking behavior across prompts. Across multiple instruction-tuned LLMs and benchmarks, our attack substantially increases turn count while remaining compliant. We also show that existing defenses offer limited protection against this emerging class of failures.

cs.LG↗

arXiv2025

Hard Work Does Not Always Pay Off: Poisoning Attacks on Neural Architecture Search

We study the robustness of data-centric methods to find neural network architectures, known as neural architecture search (NAS), against data poisoning. To audit this robustness, we design a poisoning framework that enables the systematic evaluation of the ability of NAS to produce architectures under data corruption. Our framework examines four off-the-shelf NAS algorithms, representing different approaches to architecture discovery, against four data poisoning attacks, including one we tailor specifically for NAS. In our evaluation with the CIFAR-10 and CIFAR-100 benchmarks, we show that NAS is \emph{seemingly} robust to data poisoning, showing marginal accuracy drops even under large poisoning budgets. However, we demonstrate that when considering NAS algorithms designed to achieve a few percentage points of accuracy gain, this expected improvement can be substantially diminished under data poisoning. We also show that the reduction varies across NAS algorithms and analyze the factors contributing to their robustness. Our findings are: (1) Training-based NAS algorithms are the least robust due to their reliance on data. (2) Training-free NAS approaches are the most robust but produce architectures that perform similarly to random selections from the search space. (3) NAS algorithms can produce architectures with improved accuracy, even when using out-of-distribution data like MNIST. We lastly discuss potential countermeasures. Our code is available at: https://github.com/ztcoalson/NAS-Robustness-to-Data-Poisoning

cs.LG↗

arXiv2025

IF-GUIDE: Influence Function-Guided Detoxification of LLMs

We study how training data contributes to the emergence of toxic behaviors in large language models. Most prior work on reducing model toxicity adopts reactive approaches, such as fine-tuning pre-trained (and potentially toxic) models to align them with human values. In contrast, we propose a proactive approach, IF-GUIDE, that leverages influence functions to identify and suppress harmful tokens in the training data. To this end, we first show that standard influence functions are ineffective at discovering harmful training records. We then present a novel adaptation that measures token-level attributions from training data to model toxicity, along with techniques for selecting toxic training documents and a learning objective that can be integrated into both pre-training and fine-tuning. Moreover, IF-GUIDE does not rely on human-preference data, which is typically required by existing alignment methods. In our evaluation, we demonstrate that IF-GUIDE substantially reduces both explicit and implicit toxicity-by up to 10$\times$ compared to uncensored models, and up to 3$\times$ compared to baseline alignment methods such as DPO and RAD-across both pre-training and fine-tuning scenarios. IF-GUIDE is computationally efficient: a billion-parameter model is not necessary for computing influence scores; a million-parameter model-with 7.5$\times$ fewer parameters-can effectively serve as a proxy for identifying harmful data. Our code is publicly available at: https://github.com/ztcoalson/IF-Guide

cs.LG↗

arXiv2025

Data Therapist: Eliciting Domain Knowledge from Subject Matter Experts Using Large Language Models

Effective data visualization requires not only technical proficiency but also a deep understanding of the domain-specific context in which data exists. This context often includes tacit knowledge about data provenance, quality, and intended use, which is rarely explicit in the dataset itself. Motivated by growing demands to surface tacit knowledge, we present the Data Therapist, a web-based system that helps domain experts externalize such implicit knowledge through a mixed-initiative process combining iterative Q&A with interactive annotation. Powered by a large language model, the system automatically analyzes user-supplied datasets, prompts users with targeted questions, and supports annotation at varying levels of granularity. The resulting structured knowledge base can inform both human and automated visualization design. A qualitative study with expert pairs from Accounting, Political Science, and Computer Security revealed recurring patterns in how expert reason about their data and highlighted opportunities for AI support to enhance visualization design.

cs.HC↗

arXiv2025

PrisonBreak: Jailbreaking Large Language Models with at Most Twenty-Five Targeted Bit-flips

We study a new vulnerability in commercial-scale safety-aligned large language models (LLMs): their refusal to generate harmful responses can be broken by flipping only a few bits in model parameters. Our attack jailbreaks billion-parameter language models with just 5 to 25 bit-flips, requiring up to 40$\times$ fewer bit flips than prior attacks on much smaller computer vision models. Unlike prompt-based jailbreaks, our method directly uncensors models in memory at runtime, enabling harmful outputs without requiring input-level modifications. Our key innovation is an efficient bit-selection algorithm that identifies critical bits for language model jailbreaks up to 20$\times$ faster than prior methods. We evaluate our attack on 10 open-source LLMs, achieving high attack success rates (ASRs) of 80-98% with minimal impact on model utility. We further demonstrate an end-to-end exploit via Rowhammer-based fault injection, reliably jailbreaking 5 models (69-91% ASR) on a GDDR6 GPU. Our analyses reveal that: (1) models with weaker post-training alignment require fewer bit-flips to jailbreak; (2) certain model components, e.g., value projection layers, are substantially more vulnerable; and (3) the attack is mechanistically different from existing jailbreak methods. We evaluate potential countermeasures and find that our attack remains effective against defenses at various stages of the LLM pipeline.

cs.CR↗

arXiv2025

Mind the Gap: Time-of-Check to Time-of-Use Vulnerabilities in LLM-Enabled Agents

Large Language Model (LLM)-enabled agents are rapidly emerging across a wide range of applications, but their deployment introduces vulnerabilities with security implications. While prior work has examined prompt-based attacks (e.g., prompt injection) and data-oriented threats (e.g., data exfiltration), time-of-check to time-of-use (TOCTOU) remain largely unexplored in this context. TOCTOU arises when an agent validates external state (e.g., a file or API response) that is later modified before use, enabling practical attacks such as malicious configuration swaps or payload injection. In this work, we present the first study of TOCTOU vulnerabilities in LLM-enabled agents. We introduce TOCTOU-Bench, a benchmark with 66 realistic user tasks designed to evaluate this class of vulnerabilities. As countermeasures, we adapt detection and mitigation techniques from systems security to this setting and propose prompt rewriting, state integrity monitoring, and tool-fusing. Our study highlights challenges unique to agentic workflows, where we achieve up to 25% detection accuracy using automated detection methods, a 3% decrease in vulnerable plan generation, and a 95% reduction in the attack window. When combining all three approaches, we reduce the TOCTOU vulnerabilities from an executed trajectory from 12% to 8%. Our findings open a new research direction at the intersection of AI safety and systems security.

cs.CR↗

arXiv2025

Harnessing Input-Adaptive Inference for Efficient VLN

An emerging paradigm in vision-and-language navigation (VLN) is the use of history-aware multi-modal transformer models. Given a language instruction, these models process observation and navigation history to predict the most appropriate action for an agent. While they have significantly improved performance, the scale of these models can be a bottleneck in practical settings with limited computational resources. In this work, we propose a novel input-adaptive navigation method to enhance VLN model efficiency. We first show that existing input-adaptive mechanisms fail to reduce computations without substantial performance degradation. To address this, we introduce three adaptive algorithms, each deployed at a different level: (1) To improve spatial efficiency, we selectively process panoramic views at each observation of an agent. (2) To improve intra-model efficiency, we propose importance-based adaptive thresholding for the early-exit methods. (3) To improve temporal efficiency, we implement a caching mechanism that prevents reprocessing of views previously seen by the agent. In evaluations on seven VLN benchmarks, we demonstrate over a 2$\times$ reduction in computation across three off-the-shelf agents in both standard and continuous environments. Our code is publicly available at https://github.com/secure-ai-systems-group/adaptive-vision-and-language-navigation.

cs.CV↗

arXiv2025

Modeling Neural Networks with Privacy Using Neural Stochastic Differential Equations

In this work, we study the feasibility of using neural ordinary differential equations (NODEs) to model systems with intrinsic privacy properties. Unlike conventional feedforward neural networks, which have unlimited expressivity and can represent arbitrary mappings between inputs and outputs, NODEs constrain their learning to the solution of a system of differential equations. We first examine whether this constraint reduces memorization and, consequently, the membership inference risks associated with NODEs. We conduct a comprehensive evaluation of NODEs under membership inference attacks and show that they exhibit twice the resistance compared to conventional models such as ResNets. By analyzing the variance in membership risks across different NODE models, we find that their limited expressivity leads to reduced overfitting to the training data. We then demonstrate, both theoretically and empirically, that membership inference risks can be further mitigated by utilizing a stochastic variant of NODEs: neural stochastic differential equations (NSDEs). We show that NSDEs are differentially-private (DP) learners that provide the same provable privacy guarantees as DPSGD, the de-facto mechanism for training private models. NSDEs are also effective in mitigating membership inference attacks, achieving risk levels comparable to private models trained with DP-SGD while offering an improved privacyutility trade-off. Moreover, we propose a drop-in-replacement strategy that efficiently integrates NSDEs into conventional feedforward architectures to enhance their privacy.

cs.CR↗

arXiv2025

SoK: Watermarking for AI-Generated Content

As the outputs of generative AI (GenAI) techniques improve in quality, it becomes increasingly challenging to distinguish them from human-created content. Watermarking schemes are a promising approach to address the problem of distinguishing between AI and human-generated content. These schemes embed hidden signals within AI-generated content to enable reliable detection. While watermarking is not a silver bullet for addressing all risks associated with GenAI, it can play a crucial role in enhancing AI safety and trustworthiness by combating misinformation and deception. This paper presents a comprehensive overview of watermarking techniques for GenAI, beginning with the need for watermarking from historical and regulatory perspectives. We formalize the definitions and desired properties of watermarking schemes and examine the key objectives and threat models for existing approaches. Practical evaluation strategies are also explored, providing insights into the development of robust watermarking techniques capable of resisting various attacks. Additionally, we review recent representative works, highlight open challenges, and discuss potential directions for this emerging field. By offering a thorough understanding of watermarking in GenAI, this work aims to guide researchers in advancing watermarking methods and applications, and support policymakers in addressing the broader implications of GenAI.

cs.CR↗

arXiv2025

Visualizationary: Automating Design Feedback for Visualization Designers using LLMs

Interactive visualization editors empower users to author visualizations without writing code, but do not provide guidance on the art and craft of effective visual communication. In this paper, we explore the potential of using an off-the-shelf large language models (LLMs) to provide actionable and customized feedback to visualization designers. Our implementation, VISUALIZATIONARY, demonstrates how ChatGPT can be used for this purpose through two key components: a preamble of visualization design guidelines and a suite of perceptual filters that extract salient metrics from a visualization image. We present findings from a longitudinal user study involving 13 visualization designers-6 novices, 4 intermediates, and 3 experts-who authored a new visualization from scratch over several days. Our results indicate that providing guidance in natural language via an LLM can aid even seasoned designers in refining their visualizations. All our supplemental materials are available at https://osf.io/v7hu8.

cs.HC↗

arXiv2025

Certified Robustness to Clean-Label Poisoning Using Diffusion Denoising

We present a certified defense to clean-label poisoning attacks under $\ell_2$-norm. These attacks work by injecting a small number of poisoning samples (e.g., 1%) that contain bounded adversarial perturbations into the training data to induce a targeted misclassification of a test-time input. Inspired by the adversarial robustness achieved by $randomized$ $smoothing$, we show how an off-the-shelf diffusion denoising model can sanitize the tampered training data. We extensively test our defense against seven clean-label poisoning attacks in both $\ell_2$ and $\ell_{\infty}$-norms and reduce their attack success to 0-16% with only a negligible drop in the test accuracy. We compare our defense with existing countermeasures against clean-label poisoning, showing that the defense reduces the attack success the most and offers the best model utility. Our results highlight the need for future work on developing stronger clean-label attacks and using our certified yet practical defense as a strong baseline to evaluate these attacks.

cs.CR↗

arXiv2025

MADCAT: Combating Malware Detection Under Concept Drift with Test-Time Adaptation

We present MADCAT, a self-supervised approach designed to address the concept drift problem in malware detection. MADCAT employs an encoder-decoder architecture and works by test-time training of the encoder on a small, balanced subset of the test-time data using a self-supervised objective. During test-time training, the model learns features that are useful for detecting both previously seen (old) data and newly arriving samples. We demonstrate the effectiveness of MADCAT in continuous Android malware detection settings. MADCAT consistently outperforms baseline methods in detection performance at test time. We also show the synergy between MADCAT and prior approaches in addressing concept drift in malware detection

cs.CR↗