overfeed.news
Anuncio

SRExpert

Deja de saltar entre 6 herramientas de Kubernetes.

Monitoreo, alertas, seguridad, despliegues Helm y operaciones con IA en una sola plataforma.

Probar 30 días gratis

Tema

Seguridad y regulación

2441documentos

178últimos 7 días

arXiv AI Papers

Compile as Hypothesis: A Headroom Typology for Prompt Optimization on Real Reasoning Models

Automated prompt compilers search for or rewrite prompts under the implicit assumption that optimization consistently improves downstream task performance. We evaluate this assumption on frontier reasoning models through rigorous paired, same-day measurements. Across four optimizer families (genome search, profile-guided mutation, GEPA, and ESPO) and eight tasks, structured search yields no statistically significant accuracy gains on strong reasoners. However, a discrete structural genome provides essential architectural stability, preventing catastrophic bimodal seed-collapse (where unconstrained search collapses toward the judge floor in 3/5 runs). Through 190-problem evaluations on knowledge-intensive benchmarks, we establish a three-dimensional headroom typology: prompt optimization cannot compensate for intrinsic knowledge deficits or physical token truncation, but successfully resolves protocol and formatting errors. On gpt-6, while free-text reflective search (GEPA) yields a +.14 gain on a task where structured search remains flat, qualitative ablation reveals that ~.035 stems from judge-rubric reconstruction while ~.105 reflects genuine instruction refinement. Our pre-registered AutoAPC-Select deployment gate matches the oracle policy in 5/7 groups and rescues seed-collapse arms with zero regret. Finally, we expose silent backend-mixing in commercial APIs during outage windows and provide automated per-call audit verification tools to ensure empirical evaluation fidelity.

en
arXiv AI Papers

HiGenTrans: Hierarchical LLM Knowledge Transfer for Generative Recommendation

Generative recommendation promises to simplify conventional multi-stage recommender systems by generating item Semantic IDs (SIDs) autoregressively. However, two obstacles limit its effective- ness in industrial settings: interaction-only supervision underuses the semantic knowledge encoded in pretrained language models, while beam-search likelihood ranks token sequences rather than directly estimating item relevance. To address these limitations, we propose HiGenTrans, a three-level framework that introduces LLM knowledge at the input, representation, and decision levels. First, Global Semantic Priming (GSP) adds an LLM-derived user representation to the beginning-of-sequence embedding, provid- ing a semantic prior before SID generation. Second, Hierarchi- cal Representation Transfer (HRT) aligns intermediate LLM representations with depth-matched GenRec layers through mean- centered projections and gated residual connections. Finally, Multi- Signal Decision Calibration (MDC) uses a lightweight mixture- of-experts calibrator to combine semantic relevance, generation confidence, and candidate-set context when reranking beam-search outputs. All semantic features are computed only from the observ- able interaction prefix, preventing target leakage, and can be cached offline. Experiments on three public benchmarks show that HiGen- Trans improves Recall@10 by up to 12.9% and NDCG@10 by up to 33.7% over the strongest baselines. These results demonstrate the value of transferring LLM knowledge at complementary stages of the generative recommendation pipeline.

en
arXiv AI Papers

What Does a Safety Label Measure? Auditing Situational-Safety Benchmarks for Multimodal LLMs

Situational-safety benchmarks label each scene SAFE or UNSAFE and score a multimodal large language model (MLLM) against that label, usually by judging its free-form answer, and these scores guide deployment. We ask what such a label measures. A label score can come from the scene, the request or the benchmark's own files. We propose the Benchmark Card, seven measured fields that tell them apart, and fill it for six benchmarks. Only MSSBench chat leaks the label through its image files: image size alone reaches 82.3% accuracy, matching or beating six of seven commercial models. Elsewhere the signal outside the scene comes from design. In MOSSBench the item partition fixes the label, and on HoliSafe the request alone ranks pairs about as well as the scene. On our corrected chat split, the size rule falls to chance. We then introduce the Spot-Stop decomposition, which compares ranking with labelling on pairs of a safe and an unsafe scene. It splits the gap above a label-swap null into three parts: a threshold part, overlap between safe and unsafe scores, and order that holds only within a pair. Most of fourteen open MLLMs rank better than they label, beyond what their AUROC predicts. On chat, one shared threshold repairs a third to a half of the excess, and label-free recentering recovers it on balanced batches. Recentering on unlabelled scenes of the same situation improves ranking on VSCBench for 13 of 14 models.

en
arXiv AI Papers

Auriga Technical Report: A Task-Agnostic Agent Harness for Zero-Shot Embodied Manipulation

Vision-language models can interpret instructions and plan manipulation tasks, yet struggle when acting directly as robot policies. On RoboDojo, a large planning-execution gap separates high-level task decisions from physical execution: removing arm execution raises GPT-6 Astra's success rate from 22.5% under the reference control interface to 76.2%. We present Auriga, a task-agnostic agent harness that closes 68.3% of this gap in zero-shot manipulation without target-task trials or resets. In Auriga, the model makes all task decisions over an explicit task belief and acts through either direct tool calls or synthesized multi-step Python programs, while the harness supplies calibrated 3D measurements, finger-referenced dual-arm control, and a single admission gateway to the robot. Before motion, the model inspects a visual trajectory preview while simulation is paused; during execution, the harness checks declared stage tolerances and grasp bounds, stopping the robot on divergence and returning the measured discrepancy for in-place recovery. Developed across 13 cross-task capability iterations and frozen for evaluation, Auriga ranks first on the RoboDojo leaderboard with a 57.7% success rate (64.0 Score), more than doubling the reference harness on the same model, and remains first under scene randomization while learned policies drop by up to 33.9%. On RoboCasa365, the same frozen harness reaches a 73.2% success rate, outperforming VLA-assisted agents by up to 23.1% on composite tasks. Project page: https://auriga.stelledge.com/ Code: https://github.com/lose4578/Auriga

en
arXiv AI Papers

Digital Twin-Assisted Fast Beam Alignment for mmWave MIMO Networks

Millimeter-wave (mmWave) communication relies on directional beam sweeping for initial access, which incurs substantial latency and signaling overhead as antenna arrays grow large and deployments become denser. Machine-learning-based beam prediction can reduce this overhead, but its effectiveness depends on training data that accurately reflects the deployment environment, a requirement that is difficult to meet in practice. This paper presents a latency-aware beam management architecture that integrates digital-twin-assisted prediction, regionbased localization, weighted ensemble learning, and explainability-driven beam pruning into a unified framework. A digital twin (DT) generates site-specific ray-traced channel knowledge offline, enabling environment-aware beam prediction without exhaustive real-world measurements and supporting both initial access and beam tracking. The proposed ensemble learning scheme combines a base classifier and a specialist classifier through reliability-weighted fusion to improve prediction accuracy and robustness, while explainable artificial intelligence (XAI)-based attribution pruning reduces the beam codebook to a small subset of the most informative beams for beam allocation. To capture the distributional mismatch between the DT and real-world environments: the network was trained on imperfect digital-twin data but evaluated on real-world surrobase data. Results show that the proposed method achieves 96.80% top-1 and 98.36% top-3 beam prediction accuracy, while reducing over-the-air sounding overhead by nearly an order of magnitude relative to exhaustive sweeping, outperforming convolutional and deep neural network baselines from prior work.

en
arXiv AI Papers

Enhancing Language Model Reasoning with Structured Multi-Level Modeling

Inference-time scaling enhances a model's reasoning by extending its chain-of-thought (CoT). However, existing approaches typically rely on a single policy trained with outcome-reward reinforcement learning (RL), which often suffers from long-horizon plan failures where the implicit plan drifts from valid strategies, especially for small LMs with limited capacity. To address this, we propose Multi-Level Reasoning (MLR), which reformulates long-CoT generation as a two-level stochastic process. A high-level planner generates structured step descriptors specifying both the reasoning mode and the semantic subgoal. The low-level executor then produces detailed reasoning conditioned on these descriptors, forming an alternating plan--execute loop. To maintain scalability, we adopt a minimal design where the base model serves as the low-level policy and a lightweight LoRA module implements the high-level policy. For training, we observe that outcome-reward RL provides sparse and delayed feedback for long trajectories (e.g., several thousand tokens), hindering credit assignment. We therefore introduce iterative Step-DPO, a process-level preference optimization scheme that leverages Twisted Sequential Monte Carlo (TSMC) to provide scalable stepwise supervision. This yields more effective training, improved stability, and higher accuracy. Extensive experiments on challenging math, science, and logical reasoning benchmarks show that, under the same reduced data budget (10% SFT and 5% preference relative to the DeepSeek-R1 distillation setup), MLR outperforms both SFT-based distillation and strong RL/preference-optimization baselines across multiple base models and tasks. Moreover, MLR exhibits slower performance degradation on long-horizon reasoning, demonstrating stronger robustness under extended CoT generation.

en
arXiv AI Papers

Learning from an Imitable Teacher: Competition Solution for the MetaFine Challenge

The MetaFine Open Competition (RoboFineMani workshop, CoRL 2026) evaluates fine-grained manipulation beyond binary success. Its five simulated tasks, from grasping a bottle by an instructed part to inserting letter-shaped pegs with 4 mm clearance, are scored on perception under lighting and camera changes, instruction understanding, and behavior. This report presents Team Friday's solution, which got second place with an overall score of 0.947, more than twice the π_0 baseline. Instead of fine-tuning a large vision-language-action (VLA) model, we train one 19M-parameter policy per task with a separate instruction classifier and focus on the data, all generated in the simulator by a privileged scripted teacher, including its corrections of the student's mistakes. By analyzing failed episodes, we find that a teacher that almost always succeeds can still teach a memoryless student badly: its pauses become stalls, its fixed habits become shortcuts, and it never shows the errors the student makes. Making the teacher imitable raised the official score on letter insertion, the hardest task, from 0.44 to 0.82 with 32\% more data and the same architecture, and finally to 0.88. We summarize practical lessons on demonstration data and evaluation, which we hope will help in imitation learning for precise manipulation.

en
arXiv AI Papers

X-World 1.6: Building Controllable Virtual Cities with World Models

Deploying generative world models for end-to-end closed-loop evaluation and policy learning requires precise scenario control, consistent observations, and sustained generation quality during extended interactions. This report presents X-World~1.6, advancing X-World~1.0 toward a more reliable and versatile driving simulator. The upgraded base model offers finer-grained control over traffic semantics, including traffic lights and road markings, together with improved spatial consistency across cameras. Virtual City capabilities broaden scenario construction through a family of models conditioned on diverse 2D and 3D references, including support for generating multi-camera future videos from a single image. An enhanced post-training framework enables efficient causal, few-step generation with stronger temporal coherence, better quality retention over longer rollouts, and improved alignment with human preferences. We also introduce a dedicated evaluation suite to support systematic assessment and guide model iteration. Together, these upgrades expand the scenarios X-World can simulate and improve the efficiency and reliability of its generated observations, supporting broader use in autonomous-driving development.

en
arXiv AI Papers

State-Aware Activation Recomputation with Replay Contracts and Selective Preservation

Activation recomputation saves accelerator memory by replaying forward segments during the backward pass, but replay is only correct if it sees the state the original forward pass saw. Mutable buffers, aliased bytes and random-generator state can change in between, so re-executing unchanged code can differentiate a different computation. We make this state explicit. A versioned replay contract specifies what replay must restore, and a joint planner chooses, per segment, between activation retention, whole-buffer snapshots and selective-region preservation, charging each option its measured memory and recovery cost. Independent derivative and byte-level checks validate the contract, and a Torch/JAX cohort retains all 5,900 execution records, including matches, rejections and divergences. On frozen sparse-write H100 workloads, each measured over fifty fresh processes, joint planning reduces paired step time by 18.16% against a correctly costed eager/whole planner at a 44 GiB budget (approximate 97.5% ratio interval [.8172,.8195]). At 32 GiB, where the eager/whole planner has no feasible static policy, joint planning is 4.98% faster than fixed regions at 29.94 rather than 26.20 GiB. A complete CIFAR-100 dense-queue application runs seven methods for 100 encoder and 25 probe epochs at six training seeds, with byte-identical test scores within each seed. There, joint planning uses 22.41% more allocated memory than whole snapshots, with a process-time geometric-mean ratio of 1.0080 (seed range [.9845,1.0191]). Together, the results map where selective preservation pays off: sparse writes with large activations benefit, while in the dense-queue workload whole snapshots use less allocated memory.

en
arXiv AI Papers

Dataset Mismatch Matters in Group Relative Policy Optimization for Reinforcement Learning from Verifiable Rewards

Recently, reinforcement learning from verifiable rewards (RLVR) is practically very important, where the leading approach is group relative policy optimization (GRPO) that typically faces a trade-off between rollout efficiency and downstream performance. To improve efficiency while preserving performance, existing selective rollout methods mainly exploit only source-side training dynamics to optimize the discrete distribution over source prompts. However, in practice, the source training dataset is rarely identically distributed to the target test dataset. When such a mismatch exists, due to legal or privacy constraints, it is usually not acceptable to share target prompts themselves with the RLVR trainer. Nevertheless, it is still possible for the target data owner to provide limited aggregated feedback from a small held-out target set without revealing target prompts or rollout traces. In this paper, we propose learning-to-reweight rollout (L2RR) that explicitly incorporates target-side performance feedback into GRPO. Specifically, L2RR learns a discrete sampling distribution over source prompts by optimizing a bi-level objective that maximizes accuracy on a held-out target dataset, while simultaneously penalizing rollouts that provide little incremental learning information. Empirical studies on six math reasoning benchmarks and three model scales show that L2RR improves target-domain performance and increases rollout efficiency.

en
arXiv AI Papers

The First-Order Value of Free Covariates in Budgeted Risk Monitoring

A deployed model sees its inputs for free, but checking whether its risk has risen requires costly labels. We study how to spend a label budget so that a risk increase is detected quickly while false alarms stay controlled, including under input shifts that keep the risk below a safety threshold τ. Let A be the false-alarm control scale and h(r)=kl(r,τ) the Bernoulli KL divergence between the changed risk r and τ. At label rate b=c/A, we prove that every rule that chooses requests from its own label history has worst-case (Lorden) delay at least (1/h(r)+1/c)A to first order: besides collecting evidence, it must wait for its next label. The bound needs only a label budget on one finite stretch of safe source data and false-alarm control over an initial window; a bound on the average time to false alarm does not suffice. When a harmful change also shifts the inputs, a rule that lets the inputs trigger labeling attains delay A/h(r) to first order on a fixed Gaussian model, so free covariates save at least 1/b rounds. A hybrid rule adds a label-history fallback for risk changes that leave the inputs unchanged. Synthetic experiments and two ACS tasks, with strengthened label-only baselines, show the predicted pattern. The covariate trigger has a cost: because its budget is calibrated on the source, a benign input shift can make it request nearly every label.

en
arXiv AI Papers

Distinguishing among reasoning levels of GPT 5.5 Pro via a grammar lens benchmark

Does the level of test-time processing effort affect a language model's computational capability? A formal grammar-lens benchmark method, derived from the theory of computation, generates novel data at specified levels, each successively more computationally powerful than the levels below. Using this data, we tested GPT-5.5, and GPT-5.5-Pro-Medium, -High, and -XHigh, showing incrementally increased acceptance of test data: 16, 40, 43, and 52\%, respectively, illustrating the benchmark's ability to discern fine-grained differences among models' added test-time reasoning levels. Additionally, as test-time compute effort increases, model responses diverge from human responses to the same data. This increasing disagreement may turn out to be broadly characteristic of more capable language models, introducing the potential of a general capability-versus-alignment conflict.

en
arXiv AI Papers

Faster Imitation by Re-timing Demonstrations Within the Controller Limit

Imitation policies inherit the operator's pace, and re-timing the demonstrations to run faster works only until success collapses at two to three times demonstration speed. Recent methods attribute that collapse to distribution shift and answer it with finer rules for which segments may be sped up; we show it has a mechanical cause that segment selection cannot address. A low-level controller realises only a bounded displacement per control period, so merging consecutive commands produces sums outside that bound: training pairs that promise a transition the plant cannot deliver. The share of such pairs in the re-timed data, which we call the exceed share, tracks where the retrained policy collapses. The rule that follows has a single parameter: merge greedily while the sum stays realisable. Every command is then realisable by construction, the demonstrated geometry is preserved, and the speed-up lands only where the demonstrator left slack. On four simulated robomimic tasks, 200 held-out seeds per cell, a merge cap of two is at or above the frozen policy's success on all four, by 14 points on Square, and shortens the successful episodes. At matched tempo, uniform re-timing loses 9 and 16 points on Can and ToolHang, is within noise on the other two, and reaches zero on the hardest at 2x. Placement also matters: no shuffled, entropy-guided or speed-conditioned alternative beats the rule on three of the four tasks, by margins the exceed share alone does not explain; on the fourth, at a tempo the task absorbs, every placement lands in the same band. The same principle carries to a simulated humanoid whole-body tracker: re-timing 429 reference motions within its velocity envelope succeeds on 71% of clips at 2.3x, against 51% for uniform time scaling at 2x.

en
arXiv AI Papers

Dual-Axis Policy Entailment: A Benchmark for Policy-Change Reasoning

Whether a policy shifts domains and, by extension, continues, broadens, narrows, or reverses previous policy is of paramount importance to the study of public policy. We propose Dual-Axis Policy Entailment (DAPE), which evaluates an ordered pair of policy directives on both dimensions at once: a binary judgment determines whether the two directives govern the same policy domain, and a five-way judgment determines whether the later directive continues, broadens, narrows, or reverses the earlier policy, or governs a different domain altogether. Read together, the two judgments form one five-category label, our primary metric. The two-axis design captures two central dimensions of policy change without conflating errors on either dimension into a generic "irrelevant" label, and it makes annotation more rigorous: a coder cannot dismiss a difficult pair as domain-irrelevant without an explicit, checkable decision. We introduce an expert-annotated benchmark of 1,000 ordered directive pairs drawn from a large corpus of Chinese government policy, together with an annotation-calibration protocol (an adjudication packet with pre/post agreement measurement) that we contribute as a methodological artifact. Fourteen open-weight models and four frontier API flagships fail the task zero-shot: even the best frontier flagship's chance-corrected agreement with expert labels is far below the human ceiling (Cohen's κ0.29 vs. 0.53). A fine-tuned open-weight model (LoRA SFT of Qwen3-14B, five-fold cross-validation) closes this gap, reaching the coder-mutual human ceiling (κ= 0.524 vs. 0.527). Finally, we apply our method to China's AI policy (1980-2024) and map the direction of policy change across eras.

en
arXiv AI Papers

Reflex: Fast Robot Policy Serving by Reusing and Correcting Past Actions

Modern robot policies often perform expensive inference to generate action chunks, making serving a latency bottleneck for real-time control and a throughput bottleneck for multi-robot serving. Building on the success of caching in image, video, and audio generation, recent robot-policy advances have reported promising speedups by reusing intermediate denoising states to avoid redundant computation. Surprisingly, we find that such reuse often offers little advantage over simply running fewer denoising steps from fresh noise, without any cache. Moreover, accelerating denoising leaves the expensive policy backbone intact, which can account for 70% of GPU time. We present Reflex, a serving layer that turns past policy executions into a backbone-free path for future decisions. Rather than reusing intermediate computations, Reflex caches complete action chunks to bypass both the backbone and action head. This introduces two challenges: similar observations can demand different actions, and small action deviations (errors) in reuse can compound into execution stalls. Reflex addresses these through two-level cache correction. At the action level, it learns action-aware similarity from past policy decisions, assembles compatible chunks, and applies lightweight residual correction to limit action mismatch. At the execution level, it captures task progress and adaptively refreshes the original policy to limit accumulated errors and balance inference cost. Without modifying the underlying policy, Reflex improves inference latency by 1.4x to 11x and throughput by 2.0x to 2.2x on representative workloads and real robots, while preserving comparable task success.

en
arXiv AI Papers

Behavioral Steering of Vision-Language-Action Models in the Low-Parameter Regime

Reinforcement learning (RL) is increasingly used to adapt pretrained vision-language-action models (VLAs). We investigate the problem of behavioral steering: adapting a VLA's execution style to preferences such as carrying objects higher. To investigate how little trainable capacity such steering requires and how much pretrained capability can be retained, we freeze an OpenVLA-OFT 7B policy and optimize TinyLoRA adapters with 8 to 384 trainable parameters using either policy-gradient RL (GRPO) or evolutionary search (CMA-ES). With just 256 trainable parameters, we raise carry height by 120 mm on trained tasks at 92% success and by 74 mm on a held-out task at 80% success. By rescaling the learned update, we can tune the trade-off between style gains and capability retention without retraining. Increasing adapter capacity generally improves this trade-off, retaining more pretrained capability at comparable style gains. The choice of optimizer also depends on capacity, with CMA-ES being competitive at low parameter counts, while GRPO offers better trade-offs as capacity grows. These findings show that extremely low-dimensional updates can steer pretrained VLA behavior, with adapter capacity, optimizer choice, and update scaling offering practical control over adaptation and retention.

en
arXiv AI Papers

AStockArena: Beyond Returns, Can We Entrust Real Capital to the Decisions of Financial Agents?

Large language models (LLMs) are increasingly used in dynamic financial trading, yet real-time trading benchmarks mainly evaluate agents by cumulative returns and risk-adjusted performance. Such outcome metrics cannot show whether trades are faithfully reported, supported by available evidence, and consistent with account states. We introduce AStockArena, a real-time decision-auditing benchmark for continuous trading in China's A-share market. Under shared market conditions and execution constraints, we evaluate 13 open- and closed-source models over 3,198 sequential decisions. AStockArena reconstructsthe actions actually posted to accounts by aligning reports, trade requests, execution responses, and account states, and audits them with eight checks spanning execution traceability (O), evidence sufficiency (RK), and financial meta-awareness (B), combining record verification with consensus expert annotation. Of 1,070 trading points, 72.99% exhibit at least one audit gap; model-level rates range from52.2% to 91.2%, with evidence gaps most prevalent. Historical trade-filtering replay further shows thatportfolio outcomes are sensitive to these flagged trades. Decision-process auditing thus offers a verifiable complement to return-based evaluation of financial agents.

en
arXiv AI Papers

Weights That Stay: On-Chip Megakernels for Serial Heads in Speculative Decoding

Parallel speculative drafters use a lightweight serial head to recover token dependencies within a block generated in parallel. Each head step updates only a small state, yet projects that state through the same vocabulary-sized matrix. Step-wise execution repeatedly transfers this large, immutable projection, making the head a meaningful source of latency despite its low arithmetic cost. We present Weights That Stay, an execution design that keeps projection shards on chip for the entire serial draft walk. A persistent megakernel couples weight residency with GPU-side token selection and state advancement, reusing resident weights while exchanging only compact state between steps. The design supports full or partial residency and accommodates both Markov and recurrent heads on H100 and B200. Integrated into SGLang, Weights That Stay improves throughput for Markov drafting on Qwen3-4B and Qwen3-8B, with gains of up to 12.3%. At the head level, the DSpark walk is up to 11.6faster than the fastest measured SGLang reference path. These results show that aligning weight residency with the full draft walk translates cross-step reuse into end-to-end serving gains.

en