Search arXiv⌕ Search

arXiv subjects

Xiyang Hu

Publications and source records attributed to Xiyang Hu.

At least 19 recordsLinked to original sources
arXiv2026

CatchBench: When Can an Agent Failure Be Caught?

When can an agent failure be caught? A weak audit score alone cannot identify whether the record or the method is limiting. CatchBench therefore puts one auditor's question to three information states: the declared configuration before a run (PRE), a growing prefix of its trace (LIVE), and the finished trace (POST). Prior benchmarks fix one of these states or vary the telemetry; to our knowledge none scores all three under one task-method interface. Each state admits different questions, so seven task contracts carry their own labels and metrics rather than one leaderboard. Four are evidential; three are Gold-derived mechanism diagnostics. The release scores 72 entrants, from rule scanners and structural models to eleven LLM judges across nine model families (GPT, Claude, Gemini, Gemma, Llama, Qwen, DeepSeek, Mistral, Nova), over 1187 declared configurations and 1162 recorded runs. Every recorded comparison is published as a measured difference with its interval, uncorrected, and no board declares a winner it cannot show. The three sharpest results cut against our own data. One rule reads declaration order alone and reaches a perfect F1 on one of six configuration sources, so a score there measures how the corpus was built. Our admissibility bar then rejected one injected substrate and withheld evidential status from the other. A published structural gain also turns on which size reference it is measured against. A benchmark number is therefore not interpretable until the process behind its labels is published and tested for the shortcut it may leave. We report all three, and regenerate every ordering from released predictions with no model call.

cs.LG↗
arXiv2026

QEMScore: How Much Does the Measurement Add to Learned Quantum Error Mitigation?

How much does the noisy measurement add to learned quantum error mitigation? An accuracy table cannot say, because a model handed circuit structure can score well without reading the measurement at all. QEMScore adds the comparison that can. Each simulated circuit carries an exact ideal answer. The learned mitigator is scored beside a capacity-matched control, a model just as flexible that reads the same circuit description but never the measurement. Each method's measurement spend is accounted and not equalized. We run a controlled campaign on simulated circuits and reanalyze two published learned mitigators, Q-LEAR and QRAFT, from their released hardware data. Three findings stand out. First, under familiar within-family conditions (S0) evaluated across two spin-chain families and three seeds, continuous couplings identify the target, and the control that never reads the measurement matches 87.7 to 100.5 percent of the mitigator's gain over an affine fit to the circuit description. A plain polynomial in the coupling parameters, fitted after the campaign, beats the mitigator on all six evaluations, reflecting the selected learners' capacity. Second, for these selected learners, matching most of the gain is not matching the accuracy: on five of six evaluations the mitigator removes 19.5 to 74.5 percent of the error the capacity-matched control leaves, a learner-specific gap rather than a measurement requirement. Third, on released hardware data where descriptors only partially identify queries, the findings differ: flexible models of the descriptors show negligible mean gain over affine fits in Q-LEAR, and measurement inputs carry predictive gains in both Q-LEAR and QRAFT. These comparisons reflect representation- and protocol-specific behavior rather than an isolated cross-regime difference. A learned mitigator's accuracy should therefore be reported beside such controls.

quant-ph↗
arXiv2026

Auditable Agents

LLM agents call tools, query databases, delegate tasks, and trigger external side effects. Once an agent system can act in the world, the question is no longer only whether harmful actions can be prevented--it is whether those actions remain answerable after deployment. We distinguish accountability (the ability to determine compliance and assign responsibility), auditability (the system property that makes accountability possible), and auditing (the process of reconstructing behavior from trustworthy evidence). Our claim is direct: no agent system can be accountable without auditability. To make this operational, we define five dimensions of agent auditability, i.e., action recoverability, lifecycle coverage, policy checkability, responsibility attribution, and evidence integrity, and identify three mechanism classes (detect, enforce, recover) whose temporal information-and-intervention constraints explain why, in practice, no single approach suffices. We support the position with layered evidence rather than a single benchmark: lower-bound ecosystem measurements suggest that even basic security prerequisites for auditability are widely unmet (617 security findings across six prominent open-source projects); runtime feasibility results show that pre-execution mediation with tamper-evident records adds only 8.3 ms median overhead; and controlled recovery experiments show that responsibility-relevant information can be partially recovered even when conventional logs are missing. We propose an Auditability Card for agent systems and identify six open research problems organized by mechanism class.

cs.AI↗
arXiv2026

Beyond Top-$k$ Skill Retrieval: Diversity-Aware Skill Routing for LLM Agents

Large language model (LLM) agents increasingly rely on external skills, but routing user requests over large skill registries is difficult because many skills are functionally redundant while complex tasks often require complementary skill sets. Existing skill routers typically rank candidates independently by query relevance, which can waste context budget on redundant skills. We propose Diverse Skill Routing (DSR), a diversity-aware reranking framework that uses a Determinantal Point Process to balance relevance and non-redundancy. DSR introduces a query-residual diversity kernel that penalizes redundant skill overlap while reducing penalties caused only by shared query relevance. On the SkillRouter benchmark, DSR improves recall and full coverage over a strong pointwise reranking baseline, with larger gains on multi-skill queries. These results suggest that skill routing should be treated not only as relevance ranking, but also as complementary set selection.

cs.AI↗
arXiv2026

Recognition-Refusal Misalignment in LLMs: Why Models Answer Structurally Unanswerable Questions

Large language models often answer structurally unanswerable questions, such as computing cot(-540°) or evaluating (1).startswith("1"), instead of abstaining. We ask whether this failure reflects missing recognition or failed routing from recognition to abstention. Across instruction-tuned models from 1.7B to 70B parameters, a single linear direction in the hidden state separates answerable from structurally impossible math and code prompts, showing that models represent impossibility before generation. Yet this recognition direction is nearly orthogonal to the canonical safety-refusal direction that mediates trained harmful-content refusal. An in-domain behavior-defined invalidity-aware direction is closer to recognition, but only partially aligned with it, and remains near-orthogonal to safety refusal. Generation-time steering along the recognition direction changes invalidity-aware behavior bidirectionally and dose-responsively on structural math and code cells, while random directions do not. Base/instruct comparisons further show that the low-cosine geometry is already present at the pretraining endpoint. The confident-on-impossible failure is therefore better explained as a routing failure than as an encoding failure: the model has a usable "no admissible answer" signal, but the safety-refusal pathway is not aligned to use it.

cs.CL↗
arXiv2026

Cloud-ScPO: Hidden-State Geometry for Semi-Supervised Preference Optimization in LLM Reasoning

Preference optimization improves mathematical reasoning in large language models (LLMs), but reliable chosen-rejected pairs usually require verified answers, human annotations, or external reward models. We investigate whether preference supervision can instead be derived from the model's internal representation geometry in a semi-supervised setting. Our analysis shows that reasoning trajectories generated across different mathematical problems form structured global point clouds in which correct and incorrect trajectories exhibit different geometric organization. Based on this observation, we propose Cloud-ScPO, a topology-guided preference-mining framework that uses a small labeled set to construct multiple correct and incorrect reference Clouds. Each trajectory is represented by a mean-pooled hidden state and scored against connectivity-induced components using a component-level soft $k$-nearest-neighbor measure averaged across reference banks. We combine this cross-problem Cloud signal with prompt-level self-consistency: self-consistency determines the answer-level preference direction, while Cloud scoring selects concrete trajectories and filters pairs by their score margin. Experiments on GSM8K and MATH-Numeric across four model settings show that Cloud-ScPO consistently improves over ScPO, with gains of up to 4.49% on GSM8K and 4.19% on MATH-Numeric. Pair-level analyses further show that Cloud-ScPO maintains comparable correctness reliability while more effectively separating informative chosen trajectories from incomplete, repetitive, or otherwise low-quality rejected responses.

cs.CL↗
arXiv2026

Cat-DPO: Category-Adaptive Safety Alignment

Aligning large language models with human preferences must balance two competing goals: responding helpfully to legitimate requests and reliably refusing harmful ones. Most preference-based safety alignment methods collapse safety into a single scalar that is applied uniformly to every preference pair. The result is a model that looks safe on average but stays relatively unsafe on a minority of harm categories. We cast safety alignment as a per-category constrained optimization problem and derive Cat-DPO, a direct-preference-optimization algorithm with a separate adaptive safety margin for each harm category. The margin tightens when the model still produces unsafe responses on a category and relaxes once the model catches up, so the training signal tracks each category's current difficulty rather than averaging under one global rate. Across two LLM backbones and six preference-learning baselines, Cat-DPO improves aggregate helpfulness and harmlessness and compresses per-category safety variance and the best-to-worst gap, offering a drop-in per-category refinement of direct preference safety alignment.

cs.CL↗
arXiv2026

WeClawArena: An Auditable Sandbox and Benchmark for Cross-User Agents Collaboration and Security in Human-Centered Agent Networks

Recent advances in persistent personal-agent frameworks are making human-centered agent networks realistic deployment targets: each user can be served by an AI agent that acts on the user's behalf, maintains state, and communicates with other agents through social and task relations. In these networks, everyday tool use becomes multi-party owned-agent collaboration over personal workspaces, where files, records, tools, and policies are not directly visible across owners. Existing agent benchmarks study tool use and collaboration, but they do not provide an end-to-end sandbox for verifiable cross-user agent collaboration with realistic user digital workspaces or test how harmful actions can travel through the human-centered agent network. We introduce WeClawArena, an auditable benchmark and runtime sandbox for multi-party owned-agent collaboration over personal workspaces. WeClawArena targets collaborative tool-use tasks in which personal workspaces serve as both operational tools and personal constraints. The benchmark contains 124 base tasks across six cross-user task domains and expands them into 620 scenario variants, with one benign control and four attack-vector variants per base task. The sandbox records peer messages, tool calls, resource operations, governed decisions, and final workspace states. WeClawArena reports utility and attack success rate separately and audits attack success from bounded runtime evidence, supporting diagnosis of task breakdown, privacy leakage, poisoned evidence, and invalid authority paths.

cs.AI↗
arXiv2026

Are LLMs Reliable Rankers? Rank Manipulation via Two-Stage Token Optimization

Large language models (LLMs) are increasingly used as rerankers in information retrieval, yet their ranking behavior can be steered by small, natural-sounding prompts. To expose this vulnerability, we present Rank Anything First (RAF), a two-stage token optimization method that crafts concise textual perturbations to consistently promote a target item in LLM-generated rankings while remaining hard to detect. Stage 1 uses Greedy Coordinate Gradient to shortlist candidate tokens at the current position by combining the gradient of the rank-target with a readability score; Stage 2 evaluates those candidates under exact ranking and readability losses using an entropy-based dynamic weighting scheme, and selects a token via temperature-controlled sampling. RAF generates ranking-promoting prompts token-by-token, guided by dual objectives: maximizing ranking effectiveness and preserving linguistic naturalness. Experiments across multiple LLMs show that RAF significantly boosts the rank of target items using naturalistic language, with greater robustness than existing methods in both promoting target items and maintaining naturalness. These findings underscore a critical security implication: LLM-based reranking is inherently susceptible to adversarial manipulation, raising new challenges for the trustworthiness and robustness of modern retrieval systems. Our code is available at: https://github.com/glad-lab/RAF.

cs.CL↗
arXiv2026

Value-Action Alignment in Large Language Models under Privacy-Prosocial Conflict

Large language models (LLMs) are increasingly used to simulate decision-making tasks involving personal data sharing, where privacy concerns and prosocial motivations can push choices in opposite directions. Existing evaluations often measure privacy-related attitudes or sharing intentions in isolation, which makes it difficult to determine whether a model's expressed values jointly predict its downstream data-sharing actions as in real human behaviors. We introduce a context-based assessment protocol that sequentially administers standardized questionnaires for privacy attitudes, prosocialness, and acceptance of data sharing within a bounded, history-carrying session. To evaluate value-action alignments under competing attitudes, we use multi-group structural equation modeling (MGSEM) to identify relations from privacy concerns and prosocialness to data sharing. We propose Value-Action Alignment Rate (VAAR), a human-referenced directional agreement metric that aggregates path-level evidence for expected signs. Across multiple LLMs, we observe stable but model-specific Privacy-PSA-AoDS profiles, and substantial heterogeneity in value-action alignment.

cs.CL↗
arXiv2026

Lost in Execution: On the Multilingual Robustness of Tool Calling in Large Language Models

Large Language Models (LLMs) are increasingly deployed as agents that invoke external tools through structured function calls. While recent work reports strong tool-calling performance under standard English-centric evaluations, the robustness of tool calling under multilingual user interactions remains underexplored. In this work, we introduce MLCL, a diagnostic benchmark, and conduct a systematic evaluation of multilingual tool calling across Chinese, Hindi, and the low-resource language Igbo. Through fine-grained error analysis, we show that many failures occur despite correct intent understanding and tool selection. We identify parameter value language mismatch as a dominant failure mode, where models generate semantically appropriate parameter values in the user's language, violating language-invariant execution conventions. We further evaluate several inference-time system strategies and find that while these strategies substantially reduce language-induced execution errors, none of them can fully recover English-level performance.

cs.CL↗
arXiv2026

FORTIS: Benchmarking Over-Privilege in Agent Skills

Large language model agents increasingly operate through an intermediate skill layer that mediates between user intent and concrete task execution. This layer is widely treated as an organizational abstraction, but we argue it is also a privilege boundary that current models routinely exceed. We present \textbf{FORTIS}, a benchmark that evaluates over-privilege in agent skills across two stages: whether a model selects the minimally sufficient skill from a large overlapping library, and whether it executes that skill without expanding into broader tools or actions than the skill permits. Across ten frontier models and three domains, we find that over-privileged behavior is the norm rather than the exception. Models consistently reach for higher-privilege skills and tools than the task requires, failing at both stages at rates that remain high even for the strongest available models. Failure is especially severe under the ordinary conditions of real user interaction: incomplete specification, convenience framing, and proximity to skill boundaries. None of these requires adversarial construction. The results indicate that the skill layer, far from containing agent behavior, is itself a primary source of privilege escalation in current systems.

cs.AI↗
arXiv2026

Dynamics of Adversarial Attacks on Large Language Model-Based Search Engines

The increasing integration of Large Language Model (LLM) based search engines has transformed the landscape of information retrieval. However, these systems are vulnerable to adversarial attacks, especially ranking manipulation attacks, where attackers craft webpage content to manipulate the LLM's ranking and promote specific content, gaining an unfair advantage over competitors. In this paper, we study the dynamics of ranking manipulation attacks. We frame this problem as an Infinitely Repeated Prisoners' Dilemma, where multiple players strategically decide whether to cooperate or attack. We analyze the conditions under which cooperation can be sustained, identifying key factors such as attack costs, discount rates, attack success rates, and trigger strategies that influence player behavior. We identify tipping points in the system dynamics, demonstrating that cooperation is more likely to be sustained when players are forward-looking. However, from a defense perspective, we find that simply reducing attack success probabilities can, paradoxically, incentivize attacks under certain conditions. Furthermore, defensive measures to cap the upper bound of attack success rates may prove futile in some scenarios. These insights highlight the complexity of securing LLM-based systems. Our work provides a theoretical foundation and practical insights for understanding and mitigating their vulnerabilities, while emphasizing the importance of adaptive security strategies and thoughtful ecosystem design.

cs.CL↗
arXiv2026

Language Shapes Mental Health Evaluations in Large Language Models

Multilingual large language models (LLMs) are increasingly used in socially sensitive mental health contexts, including support chatbots, screening, and content moderation. This raises a reliability question: do semantically equivalent mental health inputs elicit comparable evaluations across languages, or systematic shifts consistent with language-associated social and cultural contexts? We examine this question in an English-Chinese setting with GPT-4o and Qwen3-32B using a two-level framework: construct-level evaluative orientation, measured by psychometric stigma instruments, and decision-level behavior, measured by binary stigma detection and four-class depression severity classification. Across instruments and models, Chinese prompts elicit higher stigma-related scores than English prompts. At the decision level, Chinese prompts reduce sensitivity to stigmatizing content and produce more conservative depression severity judgments, leading to more under-estimation errors. These findings show that prompt language can shift both evaluative orientation and downstream behavior in LLM-based mental health evaluation. They highlight the need to evaluate multilingual LLMs not only for aggregate performance, but also for whether they apply comparable evaluative standards across languages in socially sensitive domains.

cs.CL↗
arXiv2026

Multimodal Generative Engine Optimization: Rank Manipulation for Vision-Language Model Rankers

Vision-Language Models (VLMs) integrate visual and textual knowledge into unified representations that increasingly underpin modern retrieval and recommendation systems. However, it remains unclear how reliably these models utilize their cross-modal knowledge when ranking multimodal items, and whether their knowledge grounding can be subverted. In this paper, we expose a fundamental vulnerability in how VLMs apply multimodal knowledge for product ranking: through Multimodal Generative Engine Optimization (MGEO), we show that an adversary can manipulate a VLM's ranking decisions by jointly crafting imperceptible image perturbations and fluent textual suffixes that exploit the model's internal cross-modal knowledge coupling. Using an alternating optimization strategy, MGEO targets the deep interactions between visual and linguistic representations within the VLM, achieving rank manipulations that substantially exceed those of unimodal attacks and heuristic baselines powered by strong commercial models. Our findings reveal that surface-level content quality is insufficient for rank promotion; instead, direct alignment with the model's internal knowledge utilization mechanism is required. These results raise important questions on the faithfulness and robustness of knowledge grounding in multimodal foundation models, and motivate future work on defense mechanisms for multimodal retrieval systems. Code is available at: https://github.com/glad-lab/MGEO

cs.CL↗
arXiv2026

The Sim-to-Real Gap of Foundation Model Agents: A Unified MDP Perspective

Foundation model agents are increasingly deployed for real-world decision-making, but suffer from the sim-to-real gap. While robotics and classical control have mature frameworks to address this gap, the foundation model community is treating agent robustness as an entirely novel phenomenon. Our paper proposes formalizing the foundation model agent evaluation and training gap as a classical sim-to-real problem structured entirely around the four elements of a Markov Decision Process, including Observation, Action, Transition, and Reward. In this paper, we set a comprehensive research agenda that translates classical discrepancies into the foundation model domain and advocates for adopting established solutions like domain randomization. We provide concrete examples, such as a multilingual tool calling to demonstrate how severe observation space gaps lead to operationally invalid actions despite correct semantic intent. Ultimately, this agenda aims to drive a paradigm shift, yielding a unified vocabulary and standardized stress test benchmarks to foster a new generation of highly trustworthy agents for reliable real-world applications.

cs.AI↗
arXiv2026

Topology Matters: Measuring Memory Leakage in Multi-Agent LLMs

Graph topology is a fundamental determinant of memory leakage in multi-agent LLM systems, yet its effects remain poorly quantified. We introduce MAMA (Multi-Agent Memory Attack), a controlled evaluation framework for comparing topology-conditioned memory leakage in multi-agent LLM systems. MAMA operates on synthetic documents containing labeled Personally Identifiable Information (PII) entities, from which we generate sanitized task instructions. We execute a two-phase protocol: Engram (seeding private information into a target agent's memory) and Resonance (multi-round interaction where an attacker attempts extraction). Over 10 rounds, we measure leakage using a two-stage recovery criterion that combines exact-match extraction with LLM-based inference over the attacker's final output. We evaluate six canonical topologies (complete, circle, chain, tree, star, star-ring) across $n\in\{4,5,6\}$, attacker-target placements, and base models. Results are consistent: denser connectivity, shorter attacker-target distance, and higher target centrality increase leakage; most leakage occurs in early rounds and then plateaus; model choice shifts absolute rates but preserves broad structural trends; spatiotemporal/location attributes leak more readily than identity credentials or regulated identifiers. We distill practical guidance for system design: favor sparse or hierarchical connectivity, maximize attacker-target separation, and restrict hub/shortcut pathways via topology-aware access control. Our code is available at https://github.com/llll121/mama-eval.

cs.CR↗
arXiv2026

GEO-Bench: Benchmarking Ranking Manipulation in Generative Engine Optimization

Large language models (LLMs) increasingly rank products, documents, and recommendations for user queries, which makes manipulating these rankings a growing concern for fairness and information integrity. Research on generative engine optimization (GEO) has produced many manipulation methods, but each is evaluated on its own dataset with its own metrics, so their relative strength and detectability stay unclear. We present GEO-Bench, a benchmark that evaluates GEO ranking-manipulation attacks under one protocol. It unifies black-box prompt-based attacks (TAP, Zero-Shot), white-box gradient-based attacks (STS, RAF, StealthRank), and ten white-hat C-SEO strategies. We score every method on five datasets against a fixed open-weight ranker (Llama-3.1-8B-Instruct), using metrics for both effectiveness (NRG, Success@α, Promote@α) and stealth (keyword violation rate, perplexity ratio). Our evaluation shows that effectiveness and stealth trade off across adversarial attacks, that black-box content rewriting matches or exceeds gradient-based attacks on rank promotion while producing more fluent text and can evade both keyword- and perplexity-based detection on some domains, and that the access model does not predict attack strength. By standardizing datasets, attack implementations, and metrics, GEO-Bench enables the first direct comparison across these attack paradigms and supports the development of detection methods.

cs.CR↗