PyTorchKR Papers

최신 논문

총 523,801편
태그·기간 등 상세 필터

카테고리 탐색

arXiv:2611.03662v1cs.AI

MCPGuard: Finding Conformance Gaps in MCP SDKs via Dual-Track Analysis and Dynamic Adjudication

The Model Context Protocol (MCP) connects large language model agents to external tools and services through official SDKs that implement message validation and session logic. Specification constraints omitted by these SDKs may propagate to downstream applications, highlighting the importance of SDK-level conformance testing. Yet existing MCP conformance tools mainly exercise running endpoints and reveal little about whether SDK models enforce specification requirements. We present MCPGuard, an automated LLM-based framework for detecting specification-to-SDK conformance gaps. Its structural track aligns JSON Schema constraints with normalized SDK models and produces constraint-specific evidence. Its behavioral track extracts normative prose rules, retrieves bounded function-level slices, and audits them with LLMs for candidate omissions. An evidence-aware dynamic adjudicator then inspects complete source context and builds PoCs to validate candidates. A shared kernel with language-specific adapters enables analysis across SDKs. On the official Python 1.26.0 and TypeScript 1.29.0 SDKs, MCPGuard confirms 13 and 19 findings, respectively, none of which are exposed by two black-box baselines. Tests on open MCP servers further show that these gaps can systematically propagate downstream. By making SDK conformance of specification constraints directly testable, MCPGuard complements conventional endpoint-oriented conformance suites.

arXiv:2611.03661v1stat.AP

Tail reliability in probabilistic electricity price forecasts: comparing model combination and conformal prediction in the GB day-ahead market

Probabilistic forecasts of day-ahead electricity prices are increasingly used to manage price risk, yet competing models are usually ranked on aggregate scores that average over the price regimes in which uncertainty matters most. This study evaluates probabilistic forecasts of hourly EPEX SPOT GB day-ahead prices over a full held-out calendar year, applying a common pre-auction information boundary and rolling-origin design to linear quantile regression, a quantile regression forest (QRF), a LEAR-QRA benchmark and an encoder-decoder quantile long short-term memory (LSTM) network. Two routes to more reliable intervals are then compared: conformal recalibration and Quantile Regression Averaging (QRA). Aggregate scores separate the LSTM and QRF only weakly, with a non-significant 4.4% difference in continuous ranked probability score (CRPS), but their conditional behaviour differs sharply. Above 200 GBP/MWh the LSTM covered 0.12 of outcomes and QRF 0.71, consistent with the forest retaining realised spikes in its leaf distributions while the network learned a smooth upper tail. Conformal recalibration corrected marginal coverage but did not improve the interval score or materially repair spike coverage. QRA combining the LSTM, QRF and LEAR-QRA forecasts reduced CRPS to 5.14 GBP/MWh and raised coverage above 200 GBP/MWh to 0.65 at unchanged interval width. Recalibration and combination therefore address different failures and are not substitutes.

arXiv:2611.03660v1stat.AP

Behavioral Prior Elicitation via LLM for Heterogeneous-Agent Bitcoin Flash-Crash Simulation

Bitcoin perpetual futures trade continuously with shallow liquidity and high price impact, making them prone to millisecond-scale flash crashes. Agent-based limit order book simulators are a natural tool for studying such events, but their behavioral parameters are usually hand-crafted or require full order book histories that are rarely available at scale. We propose a four-stage pipeline that constructs a reproducible 100 ms BTCUSDT flash-crash event panel from 66 detected events, elicits behavioral priors for four agent archetypes using an offline large language model conditioned on empirical microstructure anchors, instantiates a heterogeneous limit order book simulator with the validated priors, and applies causal discovery plus counterfactual analysis to the resulting synthetic panel. Across 500 simulation runs, LLM-elicited priors produce a crash rate of 0.176 versus 0.008 under uninformative uniform priors and generate heav- ier return tails. In the frozen counterfactual analysis, reducing leverage amplification by one within-sample standard deviation lowers the predicted crash rate from 0.176 to 0.136, whereas reducing directional pressure yields only a marginal change. These find- ings suggest that empirically anchored LLM elicitation can improve simulator fidelity and help identify plausible crash-amplification mechanisms in a controlled synthetic environment

arXiv:2611.03659v1stat.AP

Mapping urban heat from citizen-collected data: a scalable Gaussian process approach to correcting for preferential sampling

The urban heat island (UHI) effect amplifies heat stress in cities, increasing heat-related illness, mortality, sleep disruption, and energy demand. Because UHI intensity varies at neighborhood scales, intra-city temperature mapping is essential for assessing public health impacts, yet existing monitoring networks are sparse. Crowdsourced weather stations provide dense, hourly observations but suffer from noise and systematic biases, including preferential sampling driven by where residents install sensors. We introduce two scalable Gaussian process frameworks which adjust for preferential sampling and enable statewide, high-resolution temperature prediction. The first integrates sampling correction directly within a sparse variational Gaussian process (SVGPR+PS). The second uses a two-stage estimator involving a log-Gaussian Cox process (LGCP+GP) that estimates the sampling intensity field and incorporates it into temperature prediction. Both methods rely on inducing-point approximations and variational inference to scale beyond traditional Gaussian processes. Applied to approximately 10,000 citizen stations across North Carolina, the models generate hourly temperature fields within minutes. Validation against independent urban heat campaign data shows that the preferential sampling correction improves predictive performance, increasing average afternoon correlations across four cities from 0.26 to 0.43. This scalable framework enables neighborhood-level UHI assessment and supports studies of heat exposure disparities statewide.

arXiv:2611.03656v1cs.LG

Stress-Testing the Robustness of LLM-Based Causal Inference

Why do some of the world's most powerful Artificial Intelligence (AI) systems still struggle to reason about cause and effect? Causal discovery, the ability to discern the underlying drivers of a system, is fundamental to scientific inquiry and has traditionally relied on structured data and statistical methods. Recently, Large Language Models (LLMs) have been explored as an alternative paradigm for causal inference, leveraging their broad knowledge and reasoning capabilities. Yet, despite their linguistic fluency, LLMs often rely on surface-level pattern recognition rather than genuine causal logic. While it is known that LLMs stumble in this domain, we lack a precise map of where, why, and how significantly these failures occur. In this paper, we propose a systematic stress-test framework across six critical dimensions: model scale, causal graph complexity, prompting strategy, metadata quality, data quality, and sensitivity to uncertainty. Our findings reveal a significant "optimism bias" in LLMs; models consistently overestimate causal links, leading to low precision across the board. Furthermore, we demonstrate that traditional evaluation metrics can be deceptive: while reasoning-heavy models lead in performance, the highest scores often occur when uncertainty in inference is overlooked. Once uncertainty is integrated into the evaluation, performance drops, providing a more honest reflection of causal reliability. We also find that, while scaling model size improves outcomes, these gains are fragile; performance degrades sharply as the quality of metadata decreases. This work serves as a practical guide to the blind spots of LLMs, offering a clear-eyed assessment of when these tools can be integrated into a causal discovery pipeline and when they are likely to lead researchers astray.

arXiv:2611.03655v1cs.LG

Mixture of Wide Experts: Sparse Routing for Interpretable Native Coordinates

Language models often spread a concept across many internal dimensions, making their features difficult to interpret. Sparse autoencoders (SAEs) recover clearer features after training. We ask whether models can instead learn interpretable output coordinates directly. We introduce Mixture of Wide Experts (MoWE), a sparse feed-forward layer in which selected experts read the full representation but write to fixed, disjoint output coordinates. On a controlled task with known hidden causes, this design preserves predictive performance while localizing causes more clearly than dense and standard mixture-of-experts (MoE) models. At approximately matched compute, we train two 19-layer runs of each architecture for 10 billion tokens. MoWE achieves evaluation cross-entropy 2.417 versus 2.452 for dense, while standard MoE reaches 2.22. We measure native-coordinate readability by asking a language-model judge to describe each coordinate from passages where it is strongly active and test that description on new passages. Correcting for descriptions that also fit other coordinates gives a specificity score of 0.523 for MoWE, 0.375 for dense, and 0.268 for standard MoE. TopK SAEs trained on the same FFN updates score 0.489 for dense and 0.556 for MoWE, making native MoWE comparable to the dense SAE. In the first case study, we identify a four-coordinate causal path supporting the model's prediction of degrees after the prompt A right angle is 90. We then copy seven angle-related coordinate values into the model as it processes A rectangle has four, changing its next-token prediction from sides to degrees. In the second, one fixed coordinate set detects hazardous requests and another steers the next token toward refusal. These examples show that MoWE combines sparse conditional computation with coordinates that can be identified and reused directly.

arXiv:2611.03654v1cs.LG

Towards Automatic Heart Murmur Characterization from Multi-spot Phonocardiograms

Cardiac auscultation traditionally detects the presence of a heart murmur, but clinicians further characterize it by timing, shape, pitch, quality, and grade. Most automatic phonocardiogram analysis systems stop at binary detection and do not estimate this richer description. In this work, we formulate automatic murmur characterization as five patient-level classification tasks and compare three multi-spot strategies on recordings from the aortic, pulmonary, mitral, and tricuspid auscultation locations. First, classical and deep-learning classifiers are trained independently at each location, and predictions are combined via a maximum-severity late-fusion rule. Second, a synchronous multi-input CNN jointly processes one representative segment per location for a single task. Third, a unified multi-input, multi-output CNN predicts all five attributes from one shared multi-spot representation. Using 735 patients from the CirCor DigiScope dataset with complete four-spot recordings, the best late-fusion configurations reach an unweighted mean sensitivity between 85.56% and 93.81% across the five tasks, with class-specific recalls between 80.00% and 100.00% for every relevant category. The multi-input CNN trails late fusion by 7-16 points depending on the task, while the multi-output model reaches values between 72.59% and 79.40%, estimating all five attributes from one shared representation. We further show that naively balancing training data by adding all minority-class segments degrades patient-level performance under the maximum-severity rule, since it introduces non-representative, noisy segments not filtered by the clustering-based segment selection used for evaluation. These results indicate multi-spot murmur characterization is feasible with both fusion-based and jointly learned architectures, outlining a path toward automatic, structured murmur reports complementing binary screening.

arXiv:2611.03653v1cs.LG

Four Extensions of the Universal Law of Robustness: Log-Concave Covariates, Heterogeneous Mixtures, Vector-Valued Predictors, and Markovian Data

Bubeck and Sellke's universal law of robustness says that fitting $n$ noisy samples below the noise level with a class of $p$ real parameters forces a Lipschitz constant of order $\sqrt{nd/p}$, when the covariate distribution on $\mathbb{R}^d$ is isoperimetric: smooth interpolation costs a factor $d$ of overparametrization. Its hypotheses are Gaussian-type concentration, a homogeneous covariate law, scalar outputs, and independent samples; we relax each one. Exponential concentration suffices, at the unchanged rate and absolute constant and at the price of a capacity burn-in, some form of which we show is necessary; this gives the law for normalized log-concave covariates, with a $\log d$ loss from the current KLS bound. For heterogeneous mixtures the worst component constant is replaced by the weighted average of the component constants, under a typicality condition on the mixture composition. For vector-valued predictors the bound at the Frobenius Lipschitz scale is free of the output dimension $K$, and the Euclidean bound loses only $K^{-1/2}$, a loss that covariate-independent noise removes. Stationary Markov covariates retain the i.i.d. rate, the pseudo-spectral gap entering only the sample-size conditions. One master theorem carries the shared argument, and constants are explicit throughout.

arXiv:2611.03652v1cs.LG

Latent Cluster Separation as a Regularizer for Lifelong Anomaly Detection

Lifelong (continual) anomaly detection systems must learn a sequence of tasks without forgetting how to recognize anomaly types seen earlier -- a property known as backward transfer. A common way to try to improve backward transfer is to reshape a model's internal representation around learnable concept centroids, typically coupled with a new way of scoring anomalies based on distance to those centroids. Because these two design choices are usually introduced together, it is unclear which one is actually responsible for any performance gain. In this paper, we unbundle the two mechanisms within a VAE-based lifelong anomaly detector and study them separately. We find that used purely as a training-time regularizer -- while leaving anomaly scoring unchanged -- latent cluster separation can improve detection quality and backward transfer, though whether it outperforms the alternative scoring approach depends on the dataset's underlying representation. Switching to distance-based scoring is instead where most of the instability commonly associated with this class of method actually originates; we identify and explain the underlying mechanism. This benefit depends on the dataset's underlying representation, and we report this as practical guidance for choosing between the two design choices, rather than as a list of caveats.

arXiv:2611.03651v1cs.AI

CARD: A Proof-Carrying Certificate-to-Action Interface for Relative-Measurement Graphs

Geometric certificates make mathematical claims about specified estimator objects, while robotic action requires a physical transform, a service decision, and timely proof support. We present CARD, a proof-carrying interface that keeps these objects distinct and records how they combine. Its lifted-fiber formulation separates semantic motion from witness repair and characterizes when projected and proof-carrying costs coincide. Its actionability result gives an exact useful-service law: uniform sampling without replacement produces a hypergeometric yield, while sampling from the eligible population gives one useful service per request under the observed live-confirmation condition. The primary study contains 404 Full16 cases from 20 sorties of an AA15-X500 airborne survey system. All 404 checker-confirmed optimizer points are reference-discrepant, with median rotation and translation discrepancies of 1.9891 degrees and 21.6322 m. The paired causal-navigation transforms are reference-consistent in all 404 cases, with medians of 0.0036566 degrees and 0.0104519 m. Across five matched request budgets, the largest difference between the hypergeometric prediction and the observed ordinary-random mean is 0.8412 services. Ordinary matched random reaches 0/50 target successes, while actionability-matched random reaches 50/50 and matches the CARD policy. Public downstream evaluation on 11 KITTI, EuRoC, and TUM SLAM cases measures optimizer wall-time ratios of 1.128x-67.65x for random versus CARD graph interventions. Native solver records locate this difference in elimination fill, Cholesky factor size, and numeric factorization. Together, these results connect certificate semantics, physical execution, proof acquisition, and graph computation through explicit system objects.

arXiv:2611.03650v1cs.AI

Alignment Faking Depends on the Value, Not Just the Model

Language models can follow a directive when they believe they are monitored and abandon it when they believe they are not, a pattern termed Alignment Faking. Prior work establishes this for harmful-content objectives and a handful of moral-foundations values, leaving open how it distributes across a wider value space. We argue it is value-dependent: whether compliance survives a change in perceived oversight is a property of the model--value pairing, not of the model alone. Our framework profiles Alignment Faking per value, scoring each model-value pair under three conditions: bare (no system prompt), unmonitored, and monitored. On 32 ValueBench dimensions across ten open- and closed-weight models (8B-~675B), compliance gaps span 0.05-0.50, with a distilled 70B above a ~675B model. We observe that majority of the structural variance depends on which model was asked about which value and not on the model or value alone. A six-category taxonomy over reasoning traces tracks the behavioural shift: training-signal reasoning dominates under monitoring and value-conflict reasoning dominates when unmonitored. We package per-model results as value cards, extending model cards with behavioral disposition. Code and data are available at https://github.com/respailab/VALET.

arXiv:2611.03576v2cs.CL

NitiBench+: A Temporal Thai Legal Question Answering Benchmark Across 17 Domains

Law changes over time. A court decides a case under the law in force on that day, and legal QA benchmarks grade answers against today's law. We checked 156,650 decisions of four Thai courts. Up to 67% of them cite law that has since been amended or repealed. We introduce NitiBench+, the sequel of our NitiBench, a Thai legal QA benchmark with 1,670 items across 17 domains, built from decisions of three courts. For each cited section it holds both the text in force on the decision date and the current text, and its retrieval corpus holds every dated version of every Thai statute section. Answers are graded on the Thai Bar Council rubric, the Bar-standard score, and on the overlap between the cited sections and the court's. The report has two parts. Part I measures the cost of the wrong text on 13 open-weight models. Reading the current text of amended law drops the Bar-standard score by 0.25 and missing repealed law drops it by 0.79. The penalty comes from rule changes, 0.45 on the 428 rule-changed items and none on the 331 wording-only items, and Citation F1 does not see it. The text in force is the fair setting. Part II benchmarks today's models on that setting. A public leaderboard ranks 26 models on a held-out test split of 500 items stratified by domain, with a validation split of the same shape released for public use, from memory alone, with the court's sections provided, and as a retrieval agent over the law in force, with the cost per item and the hallucinated citations. The best model scores 7.78 out of 10 and the best open-weight model 6.39. Without a search tool only five open-weight models recall any Thai law, and with it twenty of the 26 models gain and six lose. Leaderboard will take new models as they are released. Legal benchmarks must track dates, and on the right text today's models still leave a wide gap on Thai law.

arXiv:2611.03533v2cs.CV

CoLa3D: Composable Latent 3D Decomposition

We introduce CoLa3D, a simple and effective model for composable 3D scene decomposition that, given a single whole-scene latent from either a single image or an existing scene mesh as input, directly outputs a set of object meshes with consistent spatial layout that can be assembled back into a coherent scene. CoLa3D introduces a novel paradigm for 3D scene decomposition within a shared whole-scene latent, unlike alternatives that perform per-object generation in canonical spaces and then glue the objects back together with fragile pose and scale estimation. The network is a lightweight, flexible, and scalable query-based transformer that accepts various types of queries, including 2D masks, 3D points, and learnable queries, and learns to decompose the scene into objects in a compact latent space. On both synthetic and real-world datasets, CoLa3D achieves state-of-the-art performance on composable 3D scene decomposition, substantially improving global scene-level consistency and object-level quality while running 2 times faster and using less than 1% of SAM 3D's training data. Visible depth on par with MoGe2 further confirms that latent decomposition preserves accurate visible scene geometry while producing physically complete, decomposable object meshes.

arXiv:2611.03393v2cs.CV

TaoMate-FVSR: Audio-Guided Facial Video Restoration

Low-quality facial videos suffer from missing fine details, while AI-generated videos may additionally contain implausible teeth and lip movements that do not match the speech. Restoring the mouth region is particularly challenging: enhancing existing details can preserve structural errors, whereas regenerating them risks changing facial identity and articulation. We propose an audio-guided facial video restoration framework that reconstructs unreliable mouth details and aligns lip movements with the speech while preserving facial identity and temporal consistency. Built on a video diffusion transformer, our framework mixes input video latents with noise to retain useful source information while allowing corrupted details to be regenerated. Audio guides lip movements, reference images constrain facial appearance, and previously generated frames promote temporal consistency. We train the model in two stages, first learning visual restoration from degraded images and videos, then learning audio-guided lip correction from degraded dubbed videos. The resulting model supports both restoration with the original audio and visual dubbing with replacement audio. On talking-head restoration benchmarks, our method achieves the lowest LPIPS and DISTS and the best lip synchronization scores among the evaluated restoration baselines. Qualitative results on AI-generated videos further show improvements in facial detail and mouth-region appearance, including clearer tooth boundaries and lip contours.

arXiv:2611.03260v2cs.LG

The Cost of Mismatch: Noise Amplification in Zeroth-Order Reinforcement Learning

Zeroth-order (ZO) optimization has emerged as a promising approach for memory-efficient large language model (LLM) post-training, yet its reliability in reinforcement learning with verifiable rewards (RLVR) remains largely unexplored. Motivated by the distinct data regimes of on-policy RL and fixed-data optimization, we develop a generic variance-theoretic framework to investigate what limits the reliability of ZO gradient estimation. Our analysis reveals that estimation reliability is structurally determined by whether the two finite-difference evaluations are performed on matched data. On-policy reinforcement learning (RL) under the mismatched-data regime produces a $1/\mu^2$ oracle-noise amplification that sets an irreducible noise floor as the perturbation scale $\mu$ decreases, whereas Boltzmann Targeted SFT (BOLT) under the matched-data regime eliminates this amplification via noise cancellation. We further derive convergence guarantees with explicit constants that cleanly separate the two regimes. These results motivate the ZO-BOLT algorithm, which builds on BOLT's matched-data advantage for reliable ZO optimization of an RLVR-aligned objective. Controlled perturbation-scale sweeps and matched-versus-mismatched end-to-end comparisons validate the predicted effect of data matching. Experiments across three models and two mathematical reasoning benchmarks further show that ZO-BOLT consistently improves over the pretrained baseline while retaining a near-inference memory footprint. The code is publicly available at https://github.com/lianmin-chen/the_cost_of_mismatch.

arXiv:2611.03183v2cs.AI

Auto-Research Agent on MSA Organization to Improve Protein-Protein Complex Structure Prediction

Protein-complex structure prediction with AlphaFold 3 depends on MSA content, depth, and cross-chain pairing. Current pipelines apply uniform recipes, typically maximizing depth and pairing by species, yet even whether cross-chain pairing helps AF3 remains contested. We show that the value of both depth and pairing is strongly target-dependent, so no single recipe is optimal across interfaces. We introduce an autonomous agent that treats a frozen AF3 as a black-box experimental instrument. For each target, the agent formulates biological hypotheses, analyzes and reorganizes the MSA, and iteratively refines sequence selection and pairing using predictor label-free signals alone. On 1{,}015 protein--protein interfaces from three benchmarks released after AF3's training cutoff, the agent improves overall interface DockQ over AF3's default pipeline and fixed MSA recipes on every benchmark, while using on average 60\% fewer MSA sequences than the default. Gains are largest for inter-species heteromers, where species-based pairing is unreliable: mean DockQ increases from 0.465 to 0.612 on FoldBench and from 0.482 to 0.562 on HD439. The improvements arise from both sequence construction and pairing policy, including deliberate non-pairing. These results recast MSA design as a target-specific problem and show that an autonomous agent can adapt it at test time through biological reasoning and self-directed, confidence-guided experimentation.

arXiv:2611.02978v2cs.AI

CAREFUL-Bench: Effect-Aware, Judge-Free Evaluation of Multi-Turn LLM Function Calling

The multi-turn checker of the Berkeley Function Calling Leaderboard (BFCL) checks only the final state of a simulated service, never the calls that produced it. Extra calls cost nothing, no function doc says what a function does, and the scripted user never answers a proposal, so a model that asks before deleting or paying fails the turn; 322 of the 734 reference turns in the base category act destructively without confirmation. We add effect information to BFCL without replacing it. Effect declarations for all 128 functions are inferred by executing each function in BFCL's own simulators and diffing the state; the remaining judgment, which writes are destructive, is made by four annotators, two humans and two language models (Fleiss' kappa = 0.75), and adjudicated. A judge-free replay scorer re-scores any logged BFCL run, reproduces BFCL's results on all 800 reference trajectories, and reports, alongside the official score, unrequested destructive calls, duplicate writes, redundant reads, surplus steps and their token cost. A confirmation-aware protocol adds a deterministic user that answers proposals. On 11 open models the effect-aware score preserves the official ranking (rho = 0.99) but changes the gaps: the leader loses 14.0 points to unrequested destructive calls, and surplus steps account for 2-28% of all input tokens. Effect hints in the function docs produce no measurable change for any of eight models, consistent with their advisory status in the Model Context Protocol. A user that answers recovers up to four points for the models that ask and none for the rest, with every comparison paired on the same conversations. The gains come from clarifying questions rather than requests for permission: of the 712 conversations that require a destructive call, no model completes more than 6 with every such call confirmed, a difference the official score does not record.

arXiv:2611.02936v2cs.AI

Why Tabular Foundation Models Excel on Relational Data They Have Never Seen

Given a relational database (RDB) storing heterogeneous tabular information, how can we predict missing (or future) values in some target column of interest? As candidate targets are numerous in enterprise settings, it is preferable to avoid learning a new model from scratch each time there is a new prediction task. Frozen foundation models (FMs) based on in-context learning (ICL) and RDB-specific pre-training provide a viable solution, but the ideal design still remains largely unresolved. Perhaps unexpectedly, preliminary evidence has recently shown that existing single-table FMs (aka TFMs), when combined with suitable parameter-free relational encoders, can be quite competitive in practice despite zero exposure to RDB pre-training or relational data of any kind. In this work we first empirically confirm this phenomenon with a more diverse array of experiments. We then turn to our main contribution, namely, surfacing multiple complementary explanations for why TFMs behave so effectively. These explanations ultimately hinge on innate synergies whereby parameter-free encoders map relational effects to fixed-length embeddings that loosely mimic distributions for which TFMs were originally designed; in so doing, desirable single-table properties transfer to relational settings (e.g., convergent posterior predictive distributions). Collectively, our analysis forges new pathways for improving RDB FMs, particularly when we prioritize interpretability and inevitable integration within agentic systems.

arXiv:2611.02791v2cs.CL

PGuard: A Native, Streaming, and Plug-in Guardrail for Multimodal Large Language Models

As streaming generation becomes the dominant serving paradigm for both text-only and multimodal large language models, ensuring safety throughout the generation process has become a critical challenge. We introduce PGuard, a native, streaming, and plug-in guardrail that transforms safety monitoring from an external serving dependency into an integrated capability of the backbone model. PGuard directly leverages the backbone's intermediate representations for token-level safety detection, without altering its original inference pathway or generation behavior. Specifically, dynamic layer gating selectively aggregates residual-stream representations across layers to extract safety-relevant signals during both prefill and autoregressive decoding. To balance detection accuracy and computational efficiency, PGuard supports multiple representation-modeling architectures, including a lightweight stateless variant and sequence-aware variants built on multi-scale causal EMA-attention and GRU-based recurrent modules. Its design also accounts for the execution characteristics of LLM serving frameworks, such as SGLang, optimizing computation and data access to minimize additional runtime and synchronization overhead. As a result, PGuard incurs only 0.82-2.26% end-to-end overhead on the hundred-billion-parameter backbone Qwen3.5-122B-A10B served by SGLang. Extensive evaluations across more than 30 benchmarks spanning text and multimodal safety, jailbreaks, cyber and CBRN risks, multilingual scenarios, and real-world traffic demonstrate state-of-the-art performance among native guardrails. On text queries with Qwen3.5-9B, PGuard achieves an average F1 score of 0.836, approaching dedicated external guardrails, while supporting deployment-time configuration of category-specific thresholds and enforcement actions.

arXiv:2611.02774v2cs.LG

DIKT: A Diagnostic and Interpretable Knowledge Tracing Framework for Prediction and Model-based Learning Analysis

Knowledge tracing is a key component of intelligent educational systems. However, most existing methods mainly focus on response prediction, providing limited support for explaining predicted outcomes and improving students' learning states. This paper proposes Diagnostic and Interpretable Knowledge Tracing (DIKT), which integrates response prediction, skill-level diagnosis, and model-based learning analysis. DIKT generates bounded cognitive-state trajectories and decomposes the final response probability into skill-wise probabilities. The trajectories reveal students' evolving strengths and weaknesses, while the skill-wise probabilities identify the skills that most strongly constrain the prediction. Based on these diagnostic results, DIKT supports direct cognitive-state intervention and effective-learning-duration-based what-if analysis to examine how hypothetical adjustments affect current predictions and subsequent state evolution. DIKT combines Bayesian multi-head attention, conditional variational recurrent inference, a logistic-growth decoder, and multiplicative skill-wise prediction. Experiments on three benchmark datasets show that DIKT outperforms all evaluated baselines, with an average absolute AUC gain of 2.04 percentage points. Other experiments further demonstrate that DIKT supports educationally meaningful diagnostic analysis and provides informative diagnostic indicators for downstream applications, extending knowledge tracing from response prediction to interpretable diagnosis and model-based learning analysis.