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。

01.
arXiv (CS.CL) 2026-06-17

Scaling Enterprise Agent Routing: Degradation, Diagnosis, and Recovery

Production LLM assistants route user requests to growing libraries of specialized tools, but how does routing accuracy degrade as the catalog scales? We study single-step routing on a 110-agent, 584-tool catalog from a deployed enterprise productivity assistant, evaluating three frontier models from 10 to 110 agents. Routing F1 on under-specified requests drops 16–23 percentage points across models. An oracle analysis decomposes the degradation into a retrieval gap (the model cannot surface the right tool) and a confusion gap (even with perfect retrieval, the oracle ceiling drops 10pp). Embedding-based shortlisting recovers +10–11pp F1 at full scale across all three models and two providers. A production annotation study (1,435 human-labeled utterances, three annotators) confirms the recovery on real traffic at +10–17pp despite 10–15pp lower absolute performance.

02.
arXiv (CS.AI) 2026-06-18

Towards Understanding What State Space Models Learn About Code

arXiv:2602.06774v2 Announce Type: replace Abstract: State Space Models (SSMs) have emerged as an efficient alternative to the Transformer architecture. Prior work shows that, when trained under comparable conditions, SSMs can match or surpass Transformers on code understanding tasks. However, their internal mechanisms remain a black box. We present the first systematic analysis of what SSM-based code models learn along with the direct comparison between SSM and Transformer models in this domain. Our analysis shows that SSMs capture syntactic and semantic structure more effectively than Transformers during pretraining but forgets certain relations during fine-tuning on some tasks. To investigate this behavior, we introduce SSM-Interpret, a frequency-domain framework that exposes a spectral shift toward short-range dependencies during fine-tuning. Guided by these findings, we propose architectural modifications that significantly improve the performance of SSM-based code model by upto +6 MRR on NLCodeSearch. This demonstrates that our analysis not only explains model behavior but also leads directly to better designs.

03.
arXiv (CS.AI) 2026-06-19

Emergent Alignment

arXiv:2606.19527v1 Announce Type: new Abstract: Can Large Language Models (LLMs) discern when their own outputs are misaligned with human ethics? And can they self-correct? We endow an LLM with a conscience step that reviews its own reasoning and outputs, and we extend the training loss with an alignment component using Direct Preference Optimization (DPO) to steer the model away from non-ethical outputs. The result is an online technique to align models in a wide range of applications: training, fine-tuning, adversarial prompting, and zero-shot learning. It does not require a weaker or stronger judge, relying instead on a frozen copy of itself. In previous work, the Emergent Misalignment scenario showed a range of emergent unethical behaviors from fine-tuning the model to hack code. Instead, we empirically show how to achieve Emergent Alignment: a single high-level introspective question steers training toward an ethical model under the same code hacking scenario.

04.
arXiv (CS.LG) 2026-06-12

Generative Modeling of Bach-Style Symbolic Music: A Comparative Study of Autoregressive, Latent-Variable, and Adversarial Approaches

arXiv:2606.13626v1 Announce Type: cross Abstract: We study generative modeling of Bach-style symbolic piano music using a shared MIDI corpus and three model families: autoregressive LSTMs with attention, latent-variable models including recurrent VAEs and vector-quantized VAEs, and generative adversarial networks. We compare their ability to model polyphonic note sequences, learn useful latent representations, and generate stylistically coherent compositions. Our experiments show that the autoregressive LSTM with attention produces the most musically coherent samples, while vector quantization helps mitigate posterior collapse and yields more structured outputs than conventional recurrent VAEs. The adversarial approach captures local pitch patterns but remains difficult to train and generalizes less reliably to Bach's style. These results highlight the relative strengths and failure modes of autoregressive, latent-variable, and adversarial approaches for symbolic music generation.

05.
arXiv (CS.CL) 2026-06-12

Layer-Resolved Optimal Transport for Hallucination Detection in NMT and Abstractive Summarization

Optimal transport (OT) has been shown to detect hallucinations in neural machine translation (NMT) by measuring the geometric distance between cross-attention distributions and a reference distribution, without any supervision. We extend this analysis to all six decoder layers of the Fairseq DE-EN model ($N=3{,}414$), showing that Wass-to-Unif and Wass-to-Data are complementary detectors specialised across hallucination types, that detection is concentrated in layers L1–L4 with L5 anti-predictive for subtler types, and that hallucinated translations lack the exploratory attention phase present in correct translations from the first decoding step. We further evaluate whether the geometric signal transfers to abstractive summarization faithfulness detection: our unsupervised OT detector on AggreFact ($N=1{,}116$) achieves $57.2\%$/$57.6\%$ balanced accuracy on CNN/XSum – above chance but substantially below supervised MiniCheck-Flan-T5-L($69.9\%$/$74.3\%$). This gap is principled: unlike NMT hallucinations, unfaithful summaries can attend correctly to source tokens while misrepresenting their content, a failure mode invisible to concentration-based OT metrics by construction. Structural experiments on T5-base confirm consistent decoder organisation across depth, with Layer~3 showing peak concentration and Layer~12 being most critical for generation quality. Together, the results establish OT on cross-attention as a reliable detector when the failure mode is source disengagement, a principled interpretability tool regardless of task, and fundamentally limited when faithfulness failures occur downstream of attention.

06.
arXiv (CS.LG) 2026-06-15

A Longitudinal Attribute-Conditioned Neural Network for Modeling Health-State Transition Probabilities in Temporally Irregular Data: The LANTERN Framework

arXiv:2606.13880v1 Announce Type: new Abstract: Accurate estimation of long-term care transition probabilities is central to disability insurance pricing, reserving, and solvency assessment. Classical actuarial multi-state models commonly rely on Markov, semi-Markov, or proportional-hazard specifications, which provide a direct connection to cohort projection but may be restrictive for irregular longitudinal health data with nonlinear aging patterns and heterogeneous covariate histories. This paper develops a well-calibrated estimator of multi-state transition probabilities for irregular longitudinal health data. The model learns from individual health history, incorporates the time elapsed between observations, and conditions transition probabilities on demographic and socioeconomic attributes. It produces a valid probability distribution over the next observed health state, with four possible states: healthy, mild disability, severe disability, and death. Individual probabilities are aggregated by age group and origin state to form transition matrices compatible with actuarial cohort projection. Using longitudinal data from the Health and Retirement Study, we compare the proposed estimator with logistic regression, gradient-boosted trees, a recurrent neural network, and a last-state persistence benchmark. The evaluation considers probabilistic accuracy, endpoint discrimination and calibration for severe disability and death, risk concentration, and transition matrix error after aggregation. The proposed estimator improves severe disability discrimination relative to logistic regression and gradient-boosted tree benchmarks, maintains strong calibration, and yields the lowest transition matrix error among the evaluated models in the held-out test analysis. Results show that a structured machine learning estimator can support long-term care transition modeling when judged by calibration and projection fidelity, beyond discrimination.

08.
arXiv (CS.AI) 2026-06-16

Thinking with Visual Grounding

arXiv:2606.16122v1 Announce Type: new Abstract: Visual thinking should not only sound right; it should show its evidence. While recent vision-language models (VLMs) can produce natural-language reasoning traces, these traces often leave the supporting image regions implicit, making them hard to verify and difficult to supervise. We introduce visually grounded thinking, a reasoning process in which models interleave natural-language thoughts with explicit point or box groundings of the visual evidence used at each step. This lets the model express intermediate reasoning in language while grounding key objects in the image regions they refer to. To train this behavior, we construct a scalable synthesis pipeline that distills correct visual reasoning traces, extracts the visual objects required by the traces, grounds them with a SAM3-based agent, and derives aligned point and box supervision from the resulting masks. We further propose grounding-aware reinforcement learning, which combines answer correctness rewards with dense grounding rewards that score whether generated object references match the correct image evidence. Across two counting benchmarks and four spatial reasoning benchmarks, adding visually grounded thinking to Gemma3-4B-IT consistently improves performance over the original model and the non-grounded thinking baseline. On spatial reasoning, the visually grounded thinking 4B models match, and in some cases surpass, Gemma3-27B-IT from the same model family. Our analysis shows that point grounding is well suited to counting, while box grounding benefits most from explicit grounding rewards on spatial tasks. Overall, our results show that VLMs think better when their intermediate thoughts are tied to the image regions that make them true.

09.
arXiv (quant-ph) 2026-06-17

On the entanglement induced by the deformation of phase-space

arXiv:2606.17587v1 Announce Type: new Abstract: Most quantum gravity theories propose that the fundamental concept of space-time is mostly compatible with quantum theory in noncommutative (NC) space. In the present paper, we revisit the notion of entanglement induced by NC deformations of phase space. The positive partial transpose (PPT) criterion for separability of bipartite Gaussian states is extended to a general class of Bopp's shift. In particular, we have considered both the position-position and momentum-momentum noncommutativity, with deformation parameters $\theta$ and $\eta$, respectively. It turns out that $\theta$ and $\eta$ induce the entanglement. We have directly applied the formalism for an anisotropic two-dimensional harmonic oscillator. Peres-Horodecki separability condition leads to a constraint equation for the parameter values of the oscillator in NC space. It turns out that the bipartite Gaussian state is almost always entangled in deformed space. To implement the theoretical idea, we provide an outline for a gedankenexperiment to identify the signature of phase-space noncommutativity, i.e., quantum gravity. In particular, the gedankenexperiment is devised to test the separability of supposedly separable Gaussian states in the usual commutative space, through the covariance matrix, which is constructed via measured output photocurrents after interaction of input Gaussian states and reference states. If the experiment shows that the supposedly separable states are actually entangled, then the entanglement is created through the intermediate background noncommutative space, which is a signature of the quantum nature of gravity.

10.
medRxiv (Medicine) 2026-06-19

Performance of family history-based colorectal cancer screening criteria by race and age at diagnosis in the Disparities and Cancer Epidemiology (DANCE) study

Importance: Family history (FH) and age are the primary criteria employed for early colorectal cancer (CRC) risk stratification. We evaluated how well these criteria identify individuals diagnosed with CRC across age and racial groups. Objective: To evaluate the performance of FH and age based screening criteria for identifying individuals with CRC, with attention to differences by race and age at diagnosis. Design, Setting, and Participants: This case control and case only analysis used data from the Disparities and Cancer Epidemiology (DANCE) cohort, a population based study of invasive CRC cases diagnosed from 2013 to 2022, recruited through the Metropolitan Detroit Cancer Surveillance System and the Louisiana Tumor Registry. Analyses included 1,158 non-Hispanic Black (NHB) and non-Hispanic White (NHW) CRC cases and 1,434 cancer-free controls from the Inflammation Health and Lung Epidemiology (INHALE) study, enrolled from the same Detroit catchment area. Data were analyzed in 2025. Exposures: Self reported cancer FH among first-degree (FD) relatives and grandparents, summarized into three FH-based screening criteria: at least one FD relative with CRC (colon early-screening criterion), any FH of Lynch syndrome related cancers, and meeting NCCN criteria for Lynch syndrome genetic testing. Main Outcomes and Measures: Proportion of cases meeting each FH based screening criterion stratified by race and age at diagnosis (

11.
arXiv (CS.AI) 2026-06-16

Estimating Mutual Information between Time Series and Temporal Event Sequences Across Diverse Analysis Tasks

arXiv:2606.01602v2 Announce Type: replace-cross Abstract: Pairwise dependence measures such as correlation and causality are fundamental to temporal data mining, yet there is still no principled and robust way to quantify dependence between heterogeneous data types, especially between continuous time series and discrete temporal event sequences. Existing approaches rely on ad hoc transformations or mutual-information estimators that are highly sensitive to quantization, repeated values, and event redundancy, leading to biased or unstable results in practice. We propose a nonparametric mutual information estimator that directly measures the dependence between time series and event sequences without data transformation, learning, or ad hoc discretization. Our method models the continuous-discrete duality of real-world time series to handle quantization and repeated-value artifacts and introduces a latent event clustering strategy to mitigate bias from event co-occurrence and redundancy. Together, these yield a robust and unified framework that bridges discrete and continuous mutual information. We evaluate the proposed estimator on four representative tasks: discrete-continuous time-delayed mutual information for causality analysis, global and local temporal repetition discovery, discrete covariate selection for time series forecasting, and continuous feature selection for classification. Experiments on synthetic and real-world datasets show consistent improvements over existing methods in accuracy, robustness, and interpretability, positioning our approach as a general-purpose dependence operator for heterogeneous temporal data, similar to Pearson correlation for homogeneous time series. Code available at: https://github.com/HaojiHu/Multimodal-Temporal-Data-Quantification

12.
arXiv (CS.CL) 2026-06-17

Analyzing and Encoding the Al-Mawrid Arabic-English Dictionary with the ISO Language Markup Framework and TEI Lex-0

This paper presents a robust methodology for the systematic digitization and encoding of the Al-Mawrid Arabic-English dictionary, transforming it from a legacy print resource into a standardized computational lexicon. Addressing a significant gap in Arabic lexical infrastructure, the study adopts a dual-standard framing that aligns the ISO Lexical Markup Framework (LMF) with the Text Encoding Initiative TEI Lex-0 guidelines. By applying an editorial view to the dictionary's macro- and microstructure, the research resolves the structural ambiguities and punctuation inconsistencies typical of 20th-century bilingual dictionaries. The methodology is grounded in an empirical analysis of the dictionary's lexical knowledge density. Drawing on a representative sample (the letter Ayn, comprising 4.6% of the total volume), the study provides scientific weight to the encoding process, demonstrating a structural parsing accuracy of 91%. Quantitative evaluation of the information extraction rules reveals high performance, with 85% precision and 98% recall for synonyms, and 88% precision for other morpho-semantic features. Beyond technical description, the paper provides a critical comparison with existing Arabic lexical resources and discusses the limitations of TEI Lex-0 when modelling specific Arabic phenomena, such as implicit "open set" semantic relations and scattered morphological cues. Furthermore, the study explores the potential for Linguistic Linked Open Data (LLOD) integration by establishing a scalable prefix-based referencing system that facilitates the resource's inclusion in the semantic web. The result is an interoperable, machine-tractable resource that provides a reproducible workflow for the retro-digitization of complex legacy bilingual lexicons within the Arabic NLP and Digital Humanities communities.

13.
arXiv (CS.AI) 2026-06-18

Agentra: A Supervisable Multi-Agent Framework for Enterprise Intrusion Response

arXiv:2606.18325v1 Announce Type: cross Abstract: Enterprise intrusion response still depends on static playbooks and analyst-driven triage, creating delay between alert generation and containment. We present Agentra, a supervisable multi-agent Intrusion Response System (IRS) framework that converts alerts from IDS, EDR, and XDR platforms into structured incident response plans grounded in MITRE ATT&CK, MITRE D3FEND, and NIST CSF 2.0. Agentra decomposes response reasoning across role-scoped agents, validates proposed plans through a bounded Planner–Validator review loop, screens retrieved threat intelligence through a Moderator security gateway, gates actions through an Action Catalog and risk score, and records decisions in an append-only audit log. We evaluate Agentra against a static OASIS CACAO v2.0 cyber-playbook baseline on a 120-event corpus drawn from ThreatHunter-Playbook, Splunk BOTSv3, and DARPA OpTC. The strongest configuration improves FP-aware IRS F1 from 0.61 to 0.84 and restores the projected harmful-action rate to the static baseline level of 0.0% after Planner-only configurations introduce unsafe overreaction. These results indicate that multi-agent response planning can improve ontology-grounded IRS coverage while preserving analyst approval and auditability.

14.
arXiv (math.PR) 2026-06-16

Sharp connectivity bounds for the vacant set of random interlacements

arXiv:2504.02777v2 Announce Type: replace Abstract: We consider percolation of the vacant set of random interlacements at intensity $u$ in dimensions three and higher, and derive lower bounds on the truncated two-point function for all values of $u>0$. These bounds are sharp up to principal exponential order for all $u$ in dimension three and all $u \neq u_\ast$ in higher dimensions, where $u_*$ refers to the critical parameter of the model, and they match the upper bounds derived in the article arXiv:2503.14497. In dimension three, our results further imply that the truncated two-point function grows at large distances $x$ at a rate that depends on $x$ only through its Euclidean norm, which offers a glimpse of the expected (Euclidean) invariance of the scaling limit at criticality. The rate function is atypical, it incurs a logarithmic correction and comes with an explicit pre-factor that converges to $0$ as the parameter $u$ approaches the critical point $u_*$ from either side. A particular challenge stems from the combined effects of lack of monotonicity due to the truncation in the super-critical phase, and the precise (rotationally invariant) controls we seek, that measure the effects of a certain "harmonic humpback" function. Among others, their derivation relies on rather fine estimates for hitting probabilities of the random walk in arbitrary direction $e$, which witness this invariance at the discrete level, and preclude straightforward applications of projection arguments.

15.
arXiv (CS.AI) 2026-06-11

Rule Taxonomy and Evolution in AI IDEs: A Mining and Survey Study

arXiv:2606.12231v1 Announce Type: cross Abstract: The adoption of AI-powered Integrated Development Environments (AI IDEs) has introduced "Rules" as a novel software artifact, allowing developers to persistently inject project-specific constraints and architectural guidelines into the context of Large Language Models (LLMs). Despite their role in aligning AI behavior with developer intent, the taxonomy, evolution, and practical impact of these rules remain largely unexplored. To bridge this gap, we conducted a mixed-methods empirical study on AI IDE rules. By mining 83 open-source projects and extracting 7,310 rules, we established a comprehensive taxonomy comprising 5 primary and 25 secondary categories. We then triangulated these artifacts with survey responses from 99 practitioners. Our analysis identified a contrast between developer priorities and actual configurations: while practitioners rate architectural constraints as highly important, rule files in repositories primarily consist of low-level workflow and code formatting constraints. Furthermore, our analysis of 1,540 rule evolution events revealed that rules are updated frequently. Repository data further indicate that rule evolution is primarily driven by constructive context expansions (29.17%) and enrichments (26.59%). In contrast, surveyed developers reported modifying rules primarily to correct AI errors (77.78%), typically by adding new negative constraints rather than editing existing ones. Finally, an artifact compliance assessment of 160 rule evolution events revealed that updating rules significantly improves the adherence of software artifacts, with the average artifact compliance rate increasing by 22.99% (from 49.14% to 72.13%) following an update. Our study provides empirical insights that can help developers optimize prompting strategies and guide tool builders in designing automated conflict-detection and context-management mechanisms for AI IDEs.

16.
arXiv (CS.AI) 2026-06-15

The Accountability Paradox: How Platform API Restrictions Undermine AI Transparency Mandates

arXiv:2505.11577v5 Announce Type: replace-cross Abstract: Recent application programming interface (API) restrictions on major social media platforms challenge compliance with the EU Digital Services Act [20], which mandates data access for algorithmic transparency. We develop a structured audit framework to assess the growing misalignment between regulatory requirements and platform implementations. Our comparative analysis of X/Twitter, Reddit, TikTok, and Meta identifies critical ``audit blind-spots'' where platform content moderation and algorithmic amplification remain inaccessible to independent verification. Our findings reveal an ``accountability paradox'': as platforms increasingly rely on AI systems, they simultaneously restrict the capacity for independent oversight. We propose targeted policy interventions aligned with the AI Risk Management Framework of the National Institute of Standards and Technology [80], emphasizing federated access models and enhanced regulatory enforcement.

17.
medRxiv (Medicine) 2026-06-16

Risk beliefs, intensive digital information and demand for a new preventative health product in public clinics: Evidence from an experiment in Zimbabwe.

Demand for preventative health care is weak in low-income settings. In a field experiment in a low-income, high-risk setting, we evaluated whether demand for a new bio-medical preventative health product, offered free at public health clinics, responds to digital feedback-based intensive information on health risks and benefits of prevention along with a clinic referral enabling access to the product. In our sample of women aged 18-24 years, we find a large correction in risk beliefs sustained six months after the intervention. Against a background of very low baseline usage, within six months we find a 5.8 percentage point increase in take up of the prevention method, a level of uptake which is very large relative to the control group. Reassuringly, there is no meaningful difference in up-take amongst baseline high- risk and low-risk individuals.

18.
arXiv (CS.LG) 2026-06-16

OptEMA: Adaptive Exponential Moving Average for Stochastic Optimization with Zero-Noise Optimality

作者:

arXiv:2603.09923v4 Announce Type: replace Abstract: Exponential moving averages (EMAs) are a central component of widely used adaptive optimizers such as Adam. However, existing analyses of Adam-style methods often yield suboptimal guarantees in the zero-noise regime, rely on open-loop parameter schedules, or require prior knowledge of smoothness constants. Motivated by these limitations, we introduce OptEMA and analyze two complementary variants: OptEMA-M, which applies an adaptive, decreasing EMA coefficient to the first moment with a fixed second-moment decay, and OptEMA-V, which swaps these roles. At the heart of these variants is a Corrected AdaGrad-Norm coefficient schedule. This formulation renders OptEMA algorithmically closed-loop and Lipschitz-free, meaning its effective stepsizes are trajectory-dependent and require no parameterization via the Lipschitz constant. Under lower-boundedness, unbiasedness, bounded variance, average smoothness, and a bounded stochastic-gradient condition used to control the adaptive normalizers, we prove that both variants achieve the unified noise-adaptive rate $\tilde{\mathcal{O}} \left(T^{-1/2}+\sigma^{1/2}T^{-1/4}\right)$ for the averaged gradient norm. In the zero-noise regime, these bounds automatically reduce to the nearly optimal deterministic rate $\widetilde{\mathcal{O}}(T^{-1/2})$ without manual hyperparameter retuning.

19.
medRxiv (Medicine) 2026-06-16

Supplementation with Arabinoxylan Dietary Fiber at Low Doses Produces Behavioral, Metabolic, and Gut Microbial Changes in Healthy, Overweight Adults: A Randomized Placebo-Controlled Trial

Background: Dietary fiber comprises a heterogeneous group of compounds with distinct physicochemical properties and biological effects. As such, functional outcomes observed for one fiber cannot be generalized to others. Some fermentable fibers, such as arabinoxylan, may exert biologically selective effects across multiple physiological domains, highlighting the need to evaluate individual ingredients for their domain-specific activity in controlled human studies. Methods: In this randomized, double-blind, parallel, 3-arm, placebo-controlled trial, healthy, overweight adults were assigned to consume one of two low doses of an arabinoxylan dietary fiber (3.5g or 5g) or placebo over the intervention period. Self-reported appetite sensations were assessed as the primary outcome using validated visual analogue scales. Secondary and exploratory endpoints included lipid parameters, gastrointestinal outcomes, mood-related measures, and gut microbiota composition and fermentation-derived metabolites. Analyses were conducted in the full analysis set and a high-compliance population to assess responses under sustained intake conditions, as per the intended dosing regimen. Results: The primary endpoint of appetite sensations did not differ between either arabinoxylan group and placebo. In contrast, evidence of microbial fermentation and selective microbiota engagement was observed. These responses occurred alongside consistent and favorable changes in lipid parameters under conditions of sustained intake, including reductions in low-density lipoprotein cholesterol and triglycerides. Additional outcomes, including gastrointestinal symptoms and mood, demonstrated domain-specific responses. Conclusion: This study demonstrates that supplementation with low doses of arabinoxylan dietary fiber elicit biologically selective, domain-specific effects across metabolic, microbial, gastrointestinal, and behavioral outcomes, particularly under conditions of sustained intake. These responses occurred independently of changes in appetite sensation, indicating that functional effects were not mediated through appetite-related pathways. Collectively, the findings highlight the ingredient's biological versatility and contextual responsiveness across physiological systems, and suggest its prebiotic potential through alignment with ISAPP's definition of a prebiotic, supporting further investigation of specific mechanistic pathways. Clinical trial registration: https://clinicaltrials.gov/study/NCT06884449, identifier: NCT06884449

20.
arXiv (CS.CL) 2026-06-11

External Experience Serving in Production LLM Systems: A Deployment-Oriented Study of Quality-Cost Trade-offs

Production LLM systems accumulate reusable operational experience, but the practical deployment issue is not merely whether such experience can help. It is how different serving strategies trade off quality against online cost under realistic constraints. Injecting external experience can improve task quality, yet it also increases prompt burden, latency, and serving pressure. We study external experience serving as a deployment-oriented quality-cost trade-off problem. We evaluate this question in a real production moderation setting, with tool-use and GPQA as supporting contrast tasks that expose different output-cost regimes. We compare no-experience baselines, random experience controls, global prompt injection, and retrieval-based selective injection, and analyze both task quality and serving cost. The results show that, once experience becomes case-dependent, selective retrieval provides a stronger operating point than unconditional global injection. They further show that retrieval quality matters more than simply increasing Top-$K$, and that the same serving policy can exhibit substantially different cost-benefit profiles across short-output and decode-heavy regimes. These findings suggest that external experience is best treated as a selective, cost-aware serving decision rather than as a universal add-on. Overall, in the settings studied here, external experience pays off only when both the serving interface and the task-specific cost structure make its quality gains worth the online cost.

21.
arXiv (CS.CL) 2026-06-17

From Observation to Intervention: A Causal Audit of Expert Importance in Mixture-of-Experts Models

Interpretability methods routinely use population-level summary statistics over observed model behaviour to license claims about the effects of targeted interventions on specific computations; in Pearl's terms, they treat rung-1 associational evidence as if it supported rung-2 interventional conclusions, a move whose validity is rarely tested. We examine one concrete instance: the use of routing statistics in Mixture-of-Experts (MoE) pruning, where utilization rates, activation norms, and routing weight distributions are treated as predictors of which experts can be removed without functional cost. A token-level interventional audit across three high-redundancy MoE architectures (OLMoE-1B-7B-0924, Qwen1.5-MoE-A2.7B, DeepSeek-V2-Lite) finds no observational metric predicts causal expert importance in any model: across all 60 metric-layer combinations effect sizes stay below Cohen's $d = 0.23$, and no metric is reliably positive under our corrected, dual-test criterion. A per-token routing weight control, run with identical $n$, rules out insufficient power, recovering a signal whose CI excludes zero at OLMoE's final MoE layer ($d = +0.231$, 95\% CI $[+0.09, +0.37]$, $p = 0.0013$). Existing pruning methods succeed in this regime not by identifying dispensable experts but because early-layer redundancy renders most selection criteria interchangeable. Our results provide an explicit counterexample to the common inferential step from population-level observational summaries to token-level interventional claims about expert importance, and illustrate how interventional audits can calibrate the evidential standards for interpretability claims.

22.
arXiv (CS.AI) 2026-06-19

How Do Instructions Shape Speech? Cross-Attention Attribution for Style-Captioned Text-to-Speech

arXiv:2606.20532v1 Announce Type: new Abstract: Style-captioned text-to-speech systems use natural language to control voice characteristics, but how individual words influence acoustic output remains unclear. Understanding this is critical for diagnosing failure modes and improving controllability in expressive TTS. We propose cross-attention attribution for speech diffusion models, adapting the DAAM framework to the speech domain for the first time, and apply it to CapSpeech-TTS. Our method extracts per-token heatmaps across 25 layers and 24 ODE steps. We analyze 3,600 (style caption, text transcript) combinations comprising 120 style captions conditioning the generation of 30 text transcripts each, revealing how caption tokens shape waveforms. Results show: (1) style tokens have lower temporal variance than content/function tokens, confirming global conditioning; (2) style attention correlates with F0 and energy; (3) style conditioning peaks in early steps and deep layers; (4) attention entropy reaches its minimum at layer 17, co-occurring with the style importance peak, indicating maximal network selectivity at the most style-critical stage. This is the first study of how natural language influences cross-attention in speech diffusion models

23.
arXiv (CS.AI) 2026-06-12

SMSR: Certified Defence Against Runtime Memory Poisoning in Persistent LLM Agent Systems

作者:

arXiv:2606.12703v1 Announce Type: cross Abstract: Retrieval-augmented generation (RAG) agents increasingly run with persistent memory that accumulates across user sessions. This creates a new attack surface: an adversary interacting only through normal channels can inject crafted memories that, once retrieved, steer the agent's responses for future users, without touching model weights or code. We call this Multi-Session Memory Poisoning (MSMP) and show that no existing defence certifies against it; static-corpus defences (RobustRAG, ReliabilityRAG) assume a fixed knowledge base, and heuristic filters are bypassed by fluent enterprise-style text. We present Signed Memory with Smoothed Retrieval (SMSR), the first defence with a certified robustness bound for this setting. Component 1 adds HMAC-SHA256 provenance at write time, blocking unsigned injection. Component 2 applies randomised memory ablation with verdict-based majority voting at query time, bounding the influence of authenticated adversaries. We prove that no provenance-free retrieval-time filter can certify against adaptive injection, derive a hypergeometric certificate for Component 2, and formalise the Consistent Minority Effect, whereby a consistent adversarial answer wins string-based voting as a numerical minority while verdict-based voting removes it. Across 15 enterprise scenarios (3,150 repeated trials), Component 1 cuts attack success from 93-100% to 0% for all unsigned variants. For an authenticated adversary with a single injection, Component 2 holds success to 8.0% (95% CI [5.8, 10.9], n=450), below the certified worst case. In an end-to-end query-only attack where the agent itself writes the poison rather than it being pre-seeded, SMSR reduces success from 65.3% to 5.3% (n=150, non-overlapping CIs) on a live agent stack. Clean-query utility is 90% (Component 1) and 85% (combined).

24.
arXiv (CS.CL) 2026-06-12

LLMs Can Better Capture Human Judgments–With the Right Prompts

Are large language models (LLMs) bad at capturing human judgment? Two commonly stated limitations are that LLMs fail to capture full distributions of responses, and that their judgments are unstable across wording variations. We demonstrate simple prompting strategies that mitigate these limitations. Across two datasets–a U.S.-representative set of 144 moral scenarios and 38 moral beliefs from the International Social Survey Programme's Family and Changing Gender Roles module covering 32 countries–we show how simple elicitation techniques help improve AI-human alignment. First, prompting models to report standard deviations and response proportions recovers the full range of human responses better than common strategies. Second, ensuring scenarios are clear to human participants–as reflected in human confusion ratings–boosts model alignment, and LLMs can track human confusion ratings. At the same time, we find that LLMs' estimates of their own error are poorly calibrated, though they can predict human variability relatively well. These results suggest that asking better questions to LLMs can yield better answers.

25.
arXiv (CS.AI) 2026-06-16

Resilient Consensus in Agentic AI

arXiv:2606.15024v1 Announce Type: cross Abstract: Large language model (LLM) agents are increasingly deployed in multi-agent systems where they must coordinate and agree on shared decisions. We ask whether classical resilient consensus theory, developed for deterministic agents, transfers to LLM agents that may behave adversarially. Framing LLM agreement as a Byzantine consensus game, we run controlled experiments on complete and general communication graphs. We find that prompted LLM agents fail to reach agreement that is achievable in principle: consensus can fail even in settings where classical theory guarantees that a convergent algorithm exists, and this failure persists across temperatures and horizons. At the same time, wrapping the agents with classical resilient consensus filters improves agreement. The benefit of filtering depends on how much robustness the underlying topology already provides. Our results suggest that classical resilient consensus theory is a useful lens for the safety of agentic AI.