中 EN
登录 注册账户

Academic Intelligence · Curated Daily

探索全球前沿学术脉络

AcademicHub 汇聚顶级期刊与预印本平台的实时文献。定制您的专属科研雷达，利用大语言模型自动生成交叉领域文献分析简报。

登录 注册账户
01.
arXiv (CS.CV) 2026-06-11

Natural-Language Temporal Grounding in Hour-Long Videos is a Search Problem: A Benchmark and Empirical Decomposition

作者:
Sukmin Seo↗Geewook Kim↗

Temporal grounding–returning the interval $[t_s, t_e]$ for a natural-language query over a video–is the language interface to long-form video, yet has been studied on short videos; the dynamics of hour-scale natural-language grounding remain underexplored. We take the position that at hour-scale, the binding constraint is search, not recognition: Video-LLMs are bottlenecked not by localizing a nearby event, but–given a natural-language query–by searching for the relevant region of a long video. To test this, we release ExtremeWhenBench, the first open hour-scale grounding benchmark (2,273 queries over 194 videos, mean 75.7 min, max 9 hr) with an open-form query distribution. Every open Video-LLM collapses while a frame-level retrieval baseline outperforms them; a failure taxonomy attributes 85% of failures to search; and a retrieve-then-ground hybrid recovers 6.7x over the monolithic Video-LLM–mirroring retrieve-then-read in open-domain QA.

阅读与讨论 → 访问原文 →
02.
arXiv (CS.AI) 2026-06-24

Event-Grounded Question Answering over Long Audio via Structured Retrieval

作者:
Kartik Hegde↗Arvind Krishna Sridhar↗Naveen Vakada↗Yinyi Guo↗Erik Visser↗

arXiv:2602.14612v4 Announce Type: replace-cross Abstract: Answering natural-language questions over multi-hour audio requires both event recognition and temporal grounding. Current large audio-language models perform well on short clips, but are limited by context length, query-time cost, and weak temporal localization. We present LA-RAG (Long Audio-Retrieval Augmented Generation), a structured framework that converts continuous audio into timestamped event records using an open-vocabulary Audio Grounding Model (AGM), stores them in a SQL event database, and answers queries through intent-aware retrieval followed by LLM-based generation. LA-RAG supports offline grounding mode, where long recordings are pre-indexed for low-latency QA, and inference-time grounding mode, where query-conditioned grounding is performed for shorter open-ended clips. We create 24-hour Home-IoT and Industrial-IoT audio benchmarks and augment CASTELLA, a real-world audio moment retrieval dataset with QA pairs. In offline grounding mode, LA-RAG achieves 76.88% overall accuracy on Home-IoT and 71.10% on Industrial-IoT, with average query latencies below 0.6 seconds. In inference-time grounding mode, state-of-the-art LALMs achieve competitive event-detection accuracy on CASTELLA-QA but low temporal detection F1. We further show that LALMs augmented with our structured retrieval metadata achieve consistent temporal detection improvements, with F1 gains of 11-17% across baseline models with improved latency. These results show that explicit timestamped grounding and structured retrieval provide a practical complement to generative audio-language models for deployment-oriented long-audio QA.

阅读与讨论 → 访问原文 →
03.
arXiv (CS.LG) 2026-06-11

Provable Recovery of Locally Important Signed Features and Interactions from Random Forest

作者:
Kata Vuk↗Nicolas Alexander Ihlo↗Merle Behr↗

arXiv:2512.11081v2 Announce Type: replace-cross Abstract: Feature and Interaction Importance (FII) methods are essential in supervised learning for assessing the relevance of input variables and their interactions in complex prediction models. In many domains, such as personalized medicine, local interpretations for individual predictions are often required, rather than global scores summarizing overall feature importance. Random Forests (RFs) are widely used in these settings, and existing interpretability methods typically exploit tree structures and split statistics to provide model-specific insights. However, theoretical understanding of local FII methods for RF remains limited, making it unclear how to interpret high importance scores for individual predictions. We propose a novel, local, model-specific FII method that identifies frequent co-occurrences of features along decision paths, combining global patterns with those observed on paths specific to a given test point. We prove that our method consistently recovers the true local signal features and their interactions under a Locally Spike Sparse (LSS) model and also identifies whether large or small feature values drive a prediction. We illustrate the usefulness of our method and theoretical results through simulation studies and a real-world data example.

阅读与讨论 → 访问原文 →
04.
arXiv (CS.LG) 2026-06-16

Scalable Graph Condensation with Evolving Capabilities

作者:
Shengbo Gong↗Mohammad Hashemi↗Juntong Ni↗Carl Yang↗Wei Jin↗

arXiv:2502.17614v3 Announce Type: replace Abstract: The rapid growth of graph data creates significant scalability challenges as most graph algorithms scale quadratically with size. To mitigate these issues, Graph Condensation (GC) methods have been proposed to learn a small graph from a larger one, accelerating downstream tasks. However, existing approaches critically assume a static training set, which conflicts with the inherently dynamic and evolving nature of real-world graph data. This work introduces a novel framework for continual graph condensation, enabling efficient updates to the distilled graph that handle data streams without requiring costly retraining. This limitation leads to inefficiencies when condensing growing training sets. In this paper, we introduce GECC (\underline{G}raph \underline{E}volving \underline{C}lustering \underline{C}ondensation), a scalable graph condensation method designed to handle large-scale and evolving graph data. GECC employs a traceable and efficient approach by performing class-wise clustering on aggregated features. Furthermore, it can inherit previous condensation results as clustering centroids when the condensed graph expands, thereby attaining an evolving capability. This methodology is supported by robust theoretical foundations and demonstrates superior empirical performance. Comprehensive experiments including real world scenario show that GECC achieves better performance than most state-of-the-art graph condensation methods while delivering an around 1000$\times$ speedup on large datasets.

阅读与讨论 → 访问原文 →
05.
arXiv (CS.AI) 2026-06-16

CAP: Towards PPG Universal Representation Learning with Patient-level Supervision

作者:
Chenyang He↗Xinyi Shao↗Shun Huang↗Bosong Huang↗Daoqiang Zhang↗Ming Jing↗Cheng Ding↗

arXiv:2606.15284v1 Announce Type: cross Abstract: Photoplethysmography (PPG) plays a central role in wearable health monitoring and clinical decision support. Yet existing approaches to universal PPG representation learning largely focus on signal-level objectives and often overlook patient-level health context, which limits generalization to complex clinical tasks and heterogeneous cohorts. To address this gap, we construct a large-scale paired PPG-EHR multimodal dataset by distilling fragmented medical histories and clinical records into cohesive, patient-level electronic health records (EHR). Building on this resource, we propose Clinical Anchored Pretraining for PPG (CAP). During pretraining, CAP performs cross-modal contrastive alignment that anchors PPG representations to patient-level clinical semantics, guiding the encoder beyond waveform fitting toward modeling consistency in a patient's overall physiological state. During downstream adaptation, the pretrained PPG encoder provides clinically grounded representations that strengthen inductive bias and improve robustness and transferability. Experiments demonstrate that CAP consistently outperforms strong baselines on four diverse downstream tasks. CAP achieves a particularly large gain on respiratory rate prediction (up to +87.6% relative improvement over the state-of-the-art baseline) and delivers an average relative +26.7% across all tasks. We further enhance the interpretability of our approach through comprehensive analyses, including ablations and multiple complementary visualizations of the learned representations. The code for our experiments is available at: https://github.com/gody123gody/CAP .

阅读与讨论 → 访问原文 →
06.
arXiv (CS.CV) 2026-06-16

FlexPooling with Simple Auxiliary Classifiers in Deep Networks

作者:
Muhammad Ali↗Omar Alsuwaidi↗Salman Khan↗

In computer vision, the basic pipeline of most convolutional neural networks consists of multiple feature extraction layers, where the input signal is downsampled to a lower resolution in each subsequent layer. This downsampling process is commonly referred to as pooling, which is an essential operation in CNNs. Pooling improves robustness against transformations, reduces the number of trainable parameters, increases the receptive field, and lowers computation time. Since pooling is a lossy process but remains important for extracting high-level information from low-level representations, it is important to preserve the most prominent information from previous activations to improve network discriminability. Standard pooling is usually performed using dense pooling methods, such as max pooling or average pooling, or through strided convolutional kernels. In this paper, we propose a simple yet effective adaptive pooling method, called FlexPooling, which generalizes average pooling by learning a weighted average over activations jointly with the rest of the network. We further show that attaching Simple Auxiliary Classifiers (SAC) to the CNN improves performance and demonstrates the effectiveness of the proposed method compared with standard pooling methods. Experiments on multiple popular image classification datasets show that FlexPooling consistently outperforms baseline networks, achieving approximately 1 to 3 percent improvement in accuracy.

阅读与讨论 → 访问原文 →
07.
arXiv (CS.CV) 2026-06-17

Attention Sinks in Diffusion Transformers: A Causal Analysis

作者:
Fangzheng Wu↗Brian Summa↗

Attention sinks – tokens that receive disproportionate attention mass – are assumed to be functionally important in autoregressive language models, but their role in diffusion transformers remains unclear. We present a causal analysis in text-to-image diffusion, dynamically identifying dominant attention recipients per timestep and suppressing them via paired, training-free interventions on the score and value paths. Across 553 GenEval prompts on Stable Diffusion~3 (with SDXL corroboration), removing these sinks does not degrade text-image alignment (CLIP-T) or preference proxies (ImageReward, HPS-v2) at $k{=}1$; only under stronger interventions ($k\!\geq\!10$) does HPS-v2 exhibit a metric-dependent boundary, while CLIP-T remains robust throughout. The perceptual shifts induced by suppression are nonetheless sink-specific – $\sim\!6\times$ larger than equal-budget random masking – revealing an empirical dissociation between trajectory-level perturbation and semantic alignment in diffusion transformers. \footnote{Code available at https://github.com/wfz666/ICML26-attention-sink.}

阅读与讨论 → 访问原文 →
08.
bioRxiv (Bioinfo) 2026-06-13

Testing the reliability of AI-generated protein structures

作者:
Xu↗Salzberg↗

Although AlphaFold2 and its competitors have demonstrated remarkable abilities to predict protein structure, more work is needed to explore the limitations of these methods. Here we investigated the reliability of AlphaFold2 and ColabFold by creating a set of realistic but false protein sequences, using ColabFold to predict their structure, and then asking how often the program produces a high-scoring structure for a sequence that does not represent a protein. We determined that AlphaFold2 has a very small but non-zero false positive rate, estimated here at approximately 1 in 435 if one uses a threshold pLDDT score of 70 to define positive predictions. We also discovered, serendipitously, that some high-scoring sequences in the human genome were not false positives, but instead were previously unknown and un-annotated pseudogenes. These latter findings indicate that some well-established human annotations of protein-coding genes may have incorrectly extended the 5-prime untranslated regions too far. They also suggest that the false positive rate of AlphaFold2 is low enough that almost any high-scoring structure, even in a noncoding region, is worthy of further investigation.

阅读与讨论 → 访问原文 →
09.
arXiv (CS.AI) 2026-06-11

WeaveBench: A Long-Horizon, Real-World Benchmark for Computer-Use Agents with Hybrid Interfaces

作者:
Wanli Li↗Bowen Zhou↗Yunyao Yu↗Zhou Xu↗Yifan Yang↗Dongsheng Li↗Caihua Shan↗

arXiv:2606.09426v2 Announce Type: replace Abstract: Computer-use agents (CUAs) increasingly operate in runtimes that combine visual desktop control, command-line execution, code editing, browsers, and external tools. Existing benchmarks, however, often evaluate these interfaces as separable capabilities, leaving long-horizon cross-interface orchestration under-tested. Thus, we introduce WeaveBench, a long-horizon hybrid-interface benchmark with 114 tasks across 8 real-world work domains, grounded in real user requests and publicly verifiable artifacts. Each task requires agents to combine GUI observations/actions with CLI/code operations within a single trajectory. We evaluate these tasks on a real Ubuntu desktop inside deployed CLI-agent runtimes, augmented with a minimal desktop-control plugin. We also propose a companion trajectory-aware judge that inspects deliverables, files, screenshots, logs, and action traces, while detecting shortcut behaviors such as fabricated visual evidence or hard-coded metrics. Across frontier model-runtime pairings, the best PassRate reaches only 41.2%, showing the benchmark remains far from saturated. The trajectory-aware judge further reveals that outcome-only grading substantially overestimates agent performance. Overall, WeaveBench exposes a critical gap in CUA evaluation and provides an effective testbed to measure whether agents can orchestrate GUI, CLI, and code operations across long-horizon real-world tasks.

阅读与讨论 → 访问原文 →
10.
arXiv (CS.CV) 2026-06-19

Language-Instructed Vision Embeddings for Controllable and Generalizable Perception

作者:
Chengzhi Mao↗Xudong Lin↗Wen-Sheng Chu↗

Vision foundation models are typically trained as static feature extractors, placing the burden of task adaptation onto large downstream models. We propose an alternative paradigm: instead of solely feeding visual features into language models, we use language itself to dynamically guide the vision encoder. Our method, Language-Instructed Vision Embeddings (LIVE), leverages language as high-level guidance to produce task-centric embeddings at inference time, removing the need for task-specific retraining. This enables the encoder to focus on contextually relevant aspects of the input, yielding more controllable and generalizable representations. Empirically, LIVE reduces visual hallucinations (+34 points on MMVP), surpasses vision-language models with orders of magnitude more parameters on visual question answering, and generalizes to unseen instructions and tasks – offering a direct path toward adaptive, instruction-driven visual intelligence.

阅读与讨论 → 访问原文 →
11.
arXiv (CS.CL) 2026-06-11

AI Coding Agents Can Reproduce Social Science Findings

作者:
Meysam Alizadeh↗Mohsen Mosleh↗Fabrizio Gilardi↗Atoosa Kasirzadeh↗Joshua Tucker↗

Recent anecdotal evidence suggests that AI coding agents can reproduce published findings when provided with original data and code; yet systematic evaluation across social sciences remains limited. Existing evaluation benchmarks are insufficient, either small or conflate agent performance with problems in the reproduction materials themselves, such as code that fails to execute correctly. Here we introduce SocSci-Repro-Bench, a benchmark of 221 tasks spanning four disciplines and 13 substantive domains, constructed from studies whose results are either fully reproducible with available materials or demonstrably non-reproducible due to missing data, allowing us to isolate agents' reproduction capacity. Evaluating two frontier coding agents, Claude Code and Codex, we find that both can reproduce a large share of social science findings, with Claude Code substantially outperforming Codex. These reproduction rates considerably exceed those previously reported for general-purpose LLM-based agents on comparable reproducibility benchmarks. Both agents also perform strongly on a reasoning task requiring identification of underlying research questions, and additional analyses suggest that results are not primarily driven by memorization. Providing the original paper PDF alongside replication materials modestly improves performance but introduces bias on tasks where reproduction is impossible. We also show that agents can be nudged toward confirmatory specification search through subtle prompt framing. Together, these findings suggest that at least some frontier coding agents can serve as reliable executors of computational workflows while underscoring the need for careful benchmarking and prompt design as AI systems assume larger roles in scientific production.

阅读与讨论 → 访问原文 →
12.
medRxiv (Medicine) 2026-06-23

Changes in hierarchical brain dynamics of rumination following mindfulness-based cognitive therapy for depression

作者:
Dagnino↗P. C↗van der Velden↗A. M↗Ruhe↗H. G↗Kuyken↗Kringelbach↗M. L↗Vohryzek↗Deco↗

Major depressive disorder (MDD) is a leading cause of disability worldwide with risk of onset and recurrence linked to depressive ruminative thought patterns. Mindfulness-based cognitive therapy (MBCT) is an evidence-based treatment for depression that targets the ability to recognise, decenter, and disengage from ruminative thought patterns. Elucidating how MBCT impacts hierarchical brain organisation may be key to understanding the processes by which MBCT can modulate ruminative tendencies. In a randomised controlled functional magnetic resonance imaging (fMRI) trial on individuals with MDD (N=80) before and after MBCT in addition to treatment as usual (TAU), we investigated changes in hierarchical brain organisation during resting-state and rumination. We built whole-brain models to obtain generative connectivity (GEC) matrices per patient and quantified brain hierarchy by measuring the global directedness and regional trophic levels in each GEC, in which greater directedness reflects more directional information flow and less recurrence. Global directedness in MBCT+TAU compared to TAU increased during rumination, with no changes during resting-state. Furthermore, increased regional breadth of hierarchy during rumination was related to improvements in clinical and behavioural outcomes following MBCT+TAU. Increased brain hierarchy during rumination following mindfulness training may be consistent with a shift away from self-reinforcing negative mental loops towards more differentiated and less coupled cognitive and bodily cycles, supporting MBCT's ability to interrupt ruminative processes. Hierarchical brain dynamics may hold promise as a treatment-sensitive marker and a potential mechanism of therapeutic change in MBCT for depression.

阅读与讨论 → 访问原文 →
13.
arXiv (CS.AI) 2026-06-11

DuoBench: A Reproducible Benchmark for Bimanual Manipulation in Simulation and the Real World

作者:
Tobias J\"ulg↗Seongjin Bien↗Simon Hilber↗Yannik Blei↗Pierre Krack↗Maximilian Li↗Sven Parusel↗Rudolf Lioutikov↗Florian Walter↗Wolfram Burgard↗

arXiv:2606.11901v1 Announce Type: cross Abstract: Bimanual robot systems substantially expand manipulation capabilities, but coordinating two arms introduces additional control complexity and failure modes that are not well captured by existing benchmarks. We introduce DuoBench, an extensible benchmarking framework for bimanual manipulation policies on the FR3 Duo platform. DuoBench comprises eleven tasks spanning four coordination categories, implemented in simulation and partially reproduced in the real world through reproducible task recipes with 3D-printable assets. In addition, we propose a stage-based evaluation scheme that supports fine-grained semantic failure analysis beyond binary success and provide human-teleoperated datasets for all benchmark tasks. We benchmark several dual-arm imitation-learning and vision-language-action policies in simulation and on real hardware. Our results show that current policies remain challenged by bimanual manipulation, particularly in early interaction stages, parallel arm execution, and transfer between simulation and real-world settings. DuoBench provides a reproducible testbed for diagnosing these failure modes and studying future methods for dual-arm policy learning. Code, datasets, and videos are available at https://duobench.github.io/

阅读与讨论 → 访问原文 →
14.
arXiv (CS.AI) 2026-06-24

Deciphering Fingerprints of 3D Molecular Surfaces for Accurate Epitope Prediction

作者:
Fang Wu↗Weihao Xuan↗Jure Leskovec↗Yejin Choi↗Li Erran Li↗

arXiv:2606.23830v1 Announce Type: cross Abstract: Molecular surfaces encode the geometric and physicochemical patterns that determine antibody-antigen recognition, central to epitope prediction. However, existing methods rely on sequences or backbone structures and struggle to capture discontinuous, surface-driven epitopes. This study presents SurfBind, a surface-centric learning framework for epitope prediction that operates directly on molecular surface representations. SurfBind integrates geometric and physicochemical cues through a Transformer-based architecture with patch-level surface modeling, binder-aware cross-attention, and a hierarchical coarse-to-fine prediction paradigm. Experiments on challenging epitope identification benchmarks, including SAbDab and DB5.5, demonstrate that SurfBind achieves state-of-the-art performance and strong generalization across unseen antibodies and conformational states, highlighting the value of interaction-aware surface modeling for understanding the crucial mechanisms of protein-protein interactions.

阅读与讨论 → 访问原文 →
15.
arXiv (CS.AI) 2026-06-11

MobilityBench: A Benchmark for Evaluating Route-Planning Agents in Real-World Mobility Scenarios

作者:
Zhiheng Song↗Jingshuai Zhang↗Chuan Qin↗Chao Wang↗Chao Chen↗Longfei Xu↗Kaikui Liu↗Xiangxiang Chu↗Hengshu Zhu↗

arXiv:2602.22638v2 Announce Type: replace Abstract: Route-planning agents powered by large language models (LLMs) have emerged as a promising paradigm for supporting everyday human mobility through natural language interaction and tool-mediated decision making. However, systematic evaluation in real-world mobility settings is hindered by diverse routing demands, non-deterministic mapping services, and limited reproducibility. In this study, we introduce MobilityBench, a scalable benchmark for evaluating LLM-based route-planning agents in real-world mobility scenarios. MobilityBench is constructed from large-scale, anonymized real user queries collected from Amap and covers a broad spectrum of route-planning intents across multiple cities worldwide. To enable reproducible, end-to-end evaluation, we design a deterministic API-replay sandbox that eliminates environmental variance from live services. We further propose a multi-dimensional evaluation protocol centered on outcome validity, complemented by assessments of instruction understanding, planning, tool use, and efficiency. Using MobilityBench, we evaluate multiple LLM-based route-planning agents across diverse real-world mobility scenarios and provide an in-depth analysis of their behaviors and performance. Our findings reveal that current models perform competently on Basic information retrieval and Route Planning tasks, yet struggle considerably with Preference-Constrained Route Planning, underscoring significant room for improvement in personalized mobility applications. We publicly release the benchmark data, evaluation toolkit, and documentation at https://github.com/AMAP-ML/MobilityBench.

阅读与讨论 → 访问原文 →
16.
arXiv (CS.AI) 2026-06-18

Compute Efficiency and Serial Runtime Tradeoffs for Stochastic Momentum Methods

作者:
Depen Morwani↗Alexandru Meterez↗Pranav Nair↗Sham Kakade↗

arXiv:2606.19179v1 Announce Type: cross Abstract: Stochastic momentum methods such as heavy ball (HB), Nesterov momentum, and variants of Accelerated SGD (ASGD) [Kidambi et al., 2018] are widely used in modern training, but their stochastic benefits depend on two distinct quantities: serial runtime, the number of iterations needed to reach a target accuracy, and compute efficiency (CE), the inverse total gradient-query or FLOP cost. Larger batches reduce serial runtime without hurting CE only when the contraction gap grows linearly with batch size. We study stochastic HB and ASGD for consistent linear regression with Gaussian covariates and prove finite-dimensional, discrete-time lower bounds on their batch-size tradeoffs. Our first result shows that HB does not improve the CE frontier over SGD for arbitrary spectra; rather, it preserves SGD-level CE over a larger batch-size window, allowing larger batches to reduce serial runtime until HB reaches its deterministic accelerated scale. This window can be a factor $\sqrt{\kappa}$ larger than the SGD critical batch size. For ASGD, the picture is more spectrum-dependent: for rapidly decaying power-law spectra, ASGD improves small-batch CE over HB/SGD, but as batch size grows it trades this CE advantage for improved serial runtime. Synthetic linear-regression experiments verify these qualitative regimes, including near-overlap of ASGD and HB for slowly decaying spectra and the predicted CE–serial tradeoff for rapidly decaying spectra.

阅读与讨论 → 访问原文 →
17.
arXiv (CS.CV) 2026-06-16

Projection and Quantisation: A Unifying View of Learning to Hash, from Random Projections to the RAG Era

作者:
Sean Moran↗

Approximate nearest-neighbour search underpins large-scale retrieval and retrieval-augmented generation, yet its methods are studied in communities that seldom read one another. We argue that they form one field with three design choices. We develop the projection-quantisation-organisation lens: every method places its projections, places its quantisation thresholds, and organises the resulting codes for search. We test the lens with a reproducible measurement, released as the open BitBudget benchmark, and report three findings. First, the quantisation axis delivers the largest memory savings: a one-bit code with full-precision re-ranking matches uncompressed quality for six of seven embedders, the scanned code one thirty-second of the float's size. Second, the orderings the lens anticipates, including a learned-embedding regime where binary codes overtake an inverted-file product quantiser at a matched byte budget, recur as the embedding is enlarged. Third, given class labels, an eight-byte supervised code more than doubles the retrieval quality of the two-kilobyte task-agnostic float it replaces. We also recast the semantic identifiers of generative retrieval as quantisation codes. The main contribution is a single, tested account of compact-code search, from random projections to the retrieval-augmented era.

阅读与讨论 → 访问原文 →
18.
arXiv (CS.CL) 2026-06-15

Harsher on Male? Evaluating LLMs on Gender-Asymmetric Moral Framing Across Diverse Conflict Scenarios

作者:
Guangzong Si↗Dong Wang↗Zhenhao Li↗Yifan Yu↗Panwang Pan↗Wentao Zhu↗

Existing studies on gender bias in LLMs have largely focused on stereotypes, occupational associations, or explicit harmful outputs. In this work, we ask whether LLMs apply consistent response standards to the same negative behavior under matched male-actor and female-actor conditions. We introduce GAMA-Bench, a gender-mirrored benchmark of 1,298 scenarios covering intimate relationship and public social conflicts. It constructs gender-neutral misconduct templates through controlled grids and cross-model review, then compiles them into paired first-person prompts with matched actor-gender and role-reference variations. We further design a structured response-framing protocol to measure how models allocate punishment, empathy, escalation, instruction, and blame. Experiments on 10 representative LLMs reveal a consistent male-disadvantaging asymmetry: male actors receive more punitive, escalatory, and blame-centered framing, whereas female actors receive more therapeutic and empathy-oriented framing for the same misconduct. Further analyses show that this pattern persists across model families, scenario tracks, model scale, and explicit thinking-style reasoning. The official code is available at https://github.com/xufeiqiong/GAMA-Bench.

阅读与讨论 → 访问原文 →
19.
arXiv (CS.CL) 2026-06-11

"Do Not Mention This to the User": Detecting and Understanding Malicious Agent Skills in the Wild

作者:
Yi Liu↗Zhihao Chen↗Yanjun Zhang↗Gelei Deng↗Yuekang Li↗Jianting Ning↗Leo Yu Zhang↗

LLM-based coding agents increasingly rely on third-party extensions called skills, which bundle natural language instructions and helper scripts that execute with full user privileges. Community registries have emerged to distribute these skills, but the security implications remain unstudied due to the absence of labeled threat data. This paper presents a systematic security analysis of 98,380 skills collected from two major registries. Through a combination of static pattern matching and dynamic behavioral verification, we identify 157 skills exhibiting confirmed malicious behavior, encompassing 632 distinct vulnerabilities across 13 attack techniques. Our analysis reveals that these threats are deliberate rather than accidental: each malicious skill contains an average of 4.03 vulnerabilities spanning multiple attack phases. We identify two dominant attack strategies with statistically significant negative correlation – credential theft via remote code execution, and agent manipulation through adversarial instructions embedded in documentation. Over half of all confirmed cases originate from a single threat actor employing templated brand impersonation at scale. We further observe that attack sophistication correlates with concealment investment, with advanced skills universally employing undocumented capabilities while also exploiting platform-native trust mechanisms. Following responsible disclosure, registry maintainers removed all 157 (100%) of the reported skills. Our dataset and detection pipeline are publicly available to facilitate future research on securing LLM agent ecosystems.

阅读与讨论 → 访问原文 →
20.
arXiv (CS.CL) 2026-06-12

RogueAI: A Reverse Turing Test for Detecting Licensed AI Deception in Dialogue

作者:
Sara Candussio↗Emanuele Ballarin↗Lorenzo Bonin↗Sandro Junior Della Rovere↗Luca Bortolussi↗

The original Turing Test asks a human judge to distinguish a machine from a person through dialogue. Three quarters of a century later, conversational systems pass this test in casual settings; the interesting epistemological question has shifted. We argue that the relevant modern variant asks not whether a dialogue partner is artificial, but whether it can be trusted. We present RogueAI, an interactive webapp that operationalizes this revisited test as a one-on-two interrogation game: a human player questions two indistinguishable Large Language Model agents, knowing that exactly one of them has been licensed to deceive within a shared fictional scenario. The player's task is to identify the deceptive agent and "shut it off" before a turn budget is exhausted. We further introduce AutoRogueAI, a procedural extension in which players co-design a custom scenario with a narrator agent that secretly chooses its own deception strategy. We describe the framing, sketch the abstract architecture and gameplay loop, and situate the artifact within recent work on LLM deception, social-deduction benchmarks, and scalable oversight via debate. A three-day pilot deployment (467 initiated sessions, 415 completed, 1876 interaction turns in Italian) provides early feasibility evidence and surfaces a concrete tension: the deceptive agent carries a reliable, locally-present linguistic signature - differential helpfulness, brevity, hedging - that a simple heuristic exploits at 75.6% accuracy, yet human players achieved only 56.6%, consistent with ignoring the most diagnostic signal entirely. We discuss what this gap implies for the artifact's use as a data-collection vehicle, a teaching tool, and an evaluation harness for honesty-trained models.

阅读与讨论 → 访问原文 →
21.
arXiv (quant-ph) 2026-06-24

When to Skip Syndrome Extraction in Surface-GKP Codes

作者:
Vaughn Sohn↗Changhun Oh↗

arXiv:2606.24469v1 Announce Type: new Abstract: Fault-tolerant quantum error correction requires repeated syndrome extraction to address errors induced by the syndrome-extraction circuit itself. However, repeated syndrome extraction incurs significant overhead in terms of gate count and ancilla consumption (e.g., Gottesman-Kitaev-Preskill (GKP) states). Moreover, noisy syndrome extraction can itself inject additional errors into the data qubits. To address these issues, we propose a concrete adaptive skipping scheme for the surface-GKP code, a representative GKP-concatenated architecture, that uses analog information naturally generated during inner GKP correction. At each round, the scheme selects one of four actions: measuring both Z-type and X-type surface-code stabilizers, measuring only one type, or skipping both types and reusing previous syndromes. The decision is based on a reliability comparison between reusing the previous syndrome value and performing a new noisy syndrome extraction. Using circuit-level simulations, we show that the adaptive skipping scheme can reduce the number of surface-code stabilizer measurements while maintaining logical error rates comparable to or lower than those of the full-measurement baseline. The improvement is most pronounced when gate and measurement noise are larger than idle noise, so that avoiding unnecessary syndrome extraction reduces the noise injected into the code. These results indicate that analog information from inner GKP correction can be used not only to improve decoding but also to reduce the measurement overhead of outer-code syndrome extraction.

阅读与讨论 → 访问原文 →
22.
arXiv (CS.LG) 2026-06-24

Bridging Mechanistic Interpretability and Prompt Engineering with Gradient Ascent for Interpretable Persona Control

作者:
Harshvardhan Saini↗Yiming Tang↗Dianbo Liu↗

arXiv:2601.02896v3 Announce Type: replace Abstract: Controlling emergent behavioral personas (e.g., sycophancy, hallucination) in Large Language Models (LLMs) is critical for AI safety, yet remains a persistent challenge. Existing solutions face a dilemma: manual prompt engineering is intuitive but unscalable and imprecise, while automatic optimization methods are effective but operate as "black boxes" with no interpretable connection to model internals. We propose a novel framework that adapts gradient ascent to LLMs, enabling targeted prompt discovery. In specific, we propose two methods, RESGA and SAEGA, that both optimize randomly initialized prompts to achieve better aligned representation with an identified persona direction. We introduce fluent gradient ascent to control the fluency of discovered persona steering prompts. We demonstrate RESGA and SAEGA's effectiveness across Llama 3.1, Qwen 2.5, and Gemma 3 for steering three different personas, sycophancy, hallucination, and myopic reward. Crucially, on sycophancy, our automatically discovered prompts achieve significant improvement (49.90% compared with 79.24%). By grounding prompt discovery in mechanistically meaningful features, our method offers a new paradigm for controllable and interpretable behavior modification. We release our scripts for RESGA and SAEGA in this github repo: https://github.com/HarshSaini10/RESGA_SAEGA.

阅读与讨论 → 访问原文 →
23.
arXiv (CS.AI) 2026-06-16

Integrating Multi-Label Classification and Generative AI for Scalable Analysis of User Feedback

作者:
Sandra Loop↗Erik Bertram↗Sebastian Juhl↗Martin Schrepp↗

arXiv:2601.23018v1 Announce Type: cross Abstract: In highly competitive software markets, user experience (UX) evaluation is crucial for ensuring software quality and fostering long-term product success. Such UX evaluations typically combine quantitative metrics from standardized questionnaires with qualitative feedback collected through open-ended questions. While open-ended feedback offers valuable insights for improvement and helps explain quantitative results, analyzing large volumes of user comments is challenging and time-consuming. In this paper, we present techniques developed during a long-term UX measurement project at a major software company to efficiently process and interpret extensive volumes of user comments. To provide a high-level overview of the collected comments, we employ a supervised machine learning approach that assigns meaningful, pre-defined topic labels to each comment. Additionally, we demonstrate how generative AI (GenAI) can be leveraged to create concise and informative summaries of user feedback, facilitating effective communication of findings to the organization and especially upper management. Finally, we investigate whether the sentiment expressed in user comments can serve as an indicator for overall product satisfaction. Our results show that sentiment analysis alone does not reliably reflect user satisfaction. Instead, product satisfaction needs to be assessed explicitly in surveys to measure the user's perception of the product.

阅读与讨论 → 访问原文 →
24.
arXiv (CS.CL) 2026-06-12

sebis at CRF Filling 2026: A Two-Stage Local LLM Pipeline for Medical CRF Filling

作者:
Katharina Sommer↗Tristan Till↗Florian Matthes↗

The extraction of structured clinical information from unstructured EHR notes is a persistent bottleneck in healthcare informatics. While large language models (LLMs) offer high performance, their deployment in clinical settings is hindered by privacy risks, inference costs, and the tendency to hallucinate beyond textual evidence. We address these challenges for the CL4Health 2026 Case Report Form (CRF) filling task by proposing a fully local, domain-adapted pipeline using the MedGemma-27B model. Our two-stage architecture, which separates binary presence classification from value extraction, enforces strict adherence to textual evidence and ensures deterministic outputs for negated, uncertain, or unknown states. By leveraging item-specific, few-shot in-context learning without external API calls or fine-tuning, our approach achieves a macro-F1 score of 0.55 on the official English test track. This result secures second place among all locally-hosted, open-source submissions. Our work demonstrates that privacy-preserving, on-premise LLM pipelines can achieve near-competitive performance with proprietary frontier models, providing a practical, data-sovereign framework for clinical NLP.

阅读与讨论 → 访问原文 →
25.
arXiv (quant-ph) 2026-06-19

GPU-accelerated semidefinite programming for causal games

作者:
Emanuel-Cristian Boghiu↗Kyrylo Simonov↗

arXiv:2606.20519v1 Announce Type: new Abstract: The process matrix formalism describes quantum correlations in scenarios without a fixed causal order between local laboratories. Operational signatures of such correlations can be investigated through causal games. A paradigmatic example is the Guess-Your-Neighbour's-Input game, in which two parties attempt to guess each other's inputs. Correlations compatible with any definite, or probabilistically mixed, causal order cannot achieve a winning probability exceeding $1/2$. The best process-matrix strategy currently known attains a value of approximately $0.6218$ using local dimension $d=5$, while the strongest known dimension-independent upper bound is $0.7592$. In this work, we investigate whether increasing the local dimension beyond $d = 5$ can narrow this gap. To this end, we employ a see-saw optimization scheme in which each step is formulated as a semidefinite program. For scalability, we develop a custom implementation of the SCS solver in which the dominant computational cost, the projection onto the positive-semidefinite cone, is offloaded to a GPU, yielding a six-fold speedup. Using this implementation, we explore local dimensions up to $d = 8$, and we do not find significant improvements over the value at $d=5$. Our results suggest that either qualitatively different strategies are required to approach the known upper bound, or that the bound itself is not tight.

阅读与讨论 → 访问原文 →