Spectral Audit of In-Context Operator Networks

Existing evaluations of neural operators and in-context operator learning rely primarily on prediction error, but accurate output prediction does not guarantee the correct local dynamical structure. A model may match solutions while exhibiting incorrect sensitivities, distorted frequency response, spurious mode coupling, or unstable tangent behavior. We introduce a Jacobian-based spectral audit for in-context operator learning. For a fixed prompt, we differentiate the network output with respect to the query function and view the resulting Jacobian as a learned tangent operator. Projecting it onto Fourier modes, we obtain a local spectral characterization of the inferred operator, including frequency-dependent gains, phase structure, and cross-mode coupling. The audit complements standard prediction metrics by testing whether the model reproduces local mechanisms of the underlying PDE operator rather than only outputs. Across benchmarks, the audit reveals distinct operator-level phenomena, including phase transport, viscosity-dependent damping, nonlinear mode coupling, and reaction--diffusion stability structure. It also detects failures partially hidden by prediction-error metrics, including high-frequency degradation, incorrect phase recovery, and prompt--operator inconsistencies. Corrupted or internally inconsistent prompts lead to degraded tangent-operator structure even when pointwise predictions remain partially accurate. Our results suggest that prediction accuracy and local operator fidelity are distinct properties of learned neural operators. Our framework also provides a diagnostic for stability, sensitivity, and operator consistency.

LL-Bench: Rethinking Low-Level Vision Evaluation in the Era of Large-Scale Generative Models

Large-scale generative models have demonstrated remarkable capabilities across image generation and editing tasks. However, their performance in low-level vision tasks, which require pixel-wise control, remains insufficiently studied. To address this gap, we introduce \textbf{LL-Bench}, a comprehensive \textbf{Benchmark} for evaluating the capabilities of large-scale generative models on \textbf{L}ow-\textbf{L}evel vision tasks. The benchmark comprises 2,469 real-world degraded images covering 16 low-level degradation tasks, and 28,919 restored images produced by 10 state-of-the-art large-scale generative models and 21 conventional restoration models, which are annotated with 152,020 expert-level pairwise human preferences and 28,334 quality scores. Built upon LL-Bench, we present a systematic diagnosis that reveals the performance boundaries and unique failure modes of large-scale generative models across diverse low-level vision tasks, compared with conventional representative restoration approaches. Moreover, we investigate the effectiveness of current quality evaluation metrics on LL-Bench, which exhibit significant discrepancy with human ratings. To better align restored-image quality assessment with human preferences, we further propose \textbf{LL-Score}, an MLLM-based evaluator that captures both restoration quality and hallucination existence. Extensive experiments demonstrate that LL-score not only outperforms existing image quality assessment metrics, but also serves as a promising reward model for training generative models on low-level vision tasks.

ROSD: Reflective On-Policy Self-Distillation for Language Model Reasoning across Domains

On-policy self-distillation (OPSD) improves the reasoning performance of large language models (LLMs) by providing dense token-level supervision for on-policy rollouts. However, existing OPSD methods often yield limited gains on in-domain reasoning and generalize poorly to out-of-domain problems. We identify two key causes: conditioning the self-teacher on a verified solution encourages imitation of training-domain reference trajectories rather than error-specific correction, and applying distillation to the full response can overwrite valid reasoning prefixes and reinforce overfitting. We propose Reflective On-policy Self-Distillation (ROSD), a framework that turns reference-solution imitation into targeted reasoning correction through reflection-guided, error-localized distillation. For each rollout, ROSD uses a self-reflector to extract a corrective idea and locate the first erroneous span. The corrective idea guides the self-teacher toward targeted supervision, while the localized error span restricts distillation to where correction is needed. This design corrects flawed reasoning while preserving valid prefixes. Experiments on multiple in-domain and out-of-domain reasoning benchmarks show that ROSD yields stronger in-domain reasoning performance overall and substantially better out-of-domain generalization than standard OPSD. Code is available at https://github.com/ZiqiZhao1/ROSD.

LongAV-Compass: Towards Unified Evaluation of Minute-Scale Audio-Visual Generation Across T2AV, I2AV, and V2AV

Audio-visual generation is rapidly advancing from short clips to minute-long content, while existing evaluation protocols remain largely confined to short-form settings. Existing benchmarks primarily focus on 5--10 second text-conditioned generation and rarely support unified evaluation across text, image, and video conditioning modalities. Moreover, they provide limited insight into how identity consistency, narrative coherence, and audio-visual alignment degrade over extended temporal horizons. To bridge this gap, we introduce LongAV-Compass, a systematic benchmark for minute-long audio-visual generation. LongAV-Compass contains 284 curated test cases spanning text-to-audio-video (T2AV), image-to-audio-video (I2AV), and video-to-audio-video (V2AV), organized by application scenario and generation complexity. The benchmark combines taxonomy-guided benchmark construction with a unified evaluation framework that integrates MLLM-assisted assessment with complementary perceptual and multimodal metrics, including DINO-v2, ArcFace, CLIP, and ImageBind. The framework evaluates more than 20 fine-grained dimensions covering within-segment quality, cross-segment consistency, global narrative coherence, semantic alignment, and audio-visual synchronization. Through experiments on 11 representative models together with human-alignment validation, LongAV-Compass provides a diagnostic testbed for analyzing the limitations of current systems in sustaining coherent, semantically aligned, and temporally consistent minute-scale audio-visual generation across diverse input modalities.

SinFormer: A Tailored Transformer for Robust Radio Frequency Fingerprint Identification

With the rapid proliferation of wireless and Internet of Things (IoT) devices, ensuring secure and reliable device identification has become a significant challenge. Traditional security techniques, such as IP or MAC address-based authentication, are susceptible to spoofing, whereas Radio Frequency Fingerprint Identification (RFFI) offers a more secure alternative by exploiting the unique hardware imperfections in devices' RF signals. In this paper, we propose a novel deep learning-based framework for RFFI that enhances both accuracy and reliability in challenging RF environments. The core of our approach is the Signal Inception Transformer (SinFormer), which leverages a specialized multi-scale self-attention mechanism to effectively capture both large-scale and fine-grained fingerprints in signals, significantly improving identification accuracy. To further enhance robustness and reliability, we introduce a two-stage training strategy that enables the model to learn general signal features and maintain performance under adverse conditions, such as low Signal-to-Noise Ratio (SNR) or channel variations. The effectiveness of the proposed method is validated using a real-world dataset. Experimental results show that the SinFormer framework consistently outperforms existing methods in accuracy and robustness across diverse and challenging scenarios.

One Algorithm, Two Goals: Dual Scoring for Parameter and Data Selection in LLM Fine-Tuning

In Large Language Model (LLM) fine-tuning, parameter and data selection are common strategies for reducing fine-tuning cost, yet they are typically driven by separate scoring mechanisms. When a parameter mask and data subset jointly determine restricted fine-tuning, this separation incurs redundant overhead and makes coordinated selection difficult. We cast parameter and data selection as two bilevel selection problems under a common validation objective and derive a shared local response-surrogate scoring rule. Under first- and second-order validation-improvement approximations, parameter importance and data utility emerge as column-wise and row-wise aggregations of a single gradient interaction matrix, yielding a closed-form row-column correspondence for co-extracting both signals. Building on this structure, we propose DualSFT (Dual-Selection Fine-Tuning), a one-shot dual-scoring algorithm that produces a parameter mask and data subset from shared gradient statistics. On 3B-9B LLMs, single-axis DualSFT variants strengthen target-task performance and stability-plasticity trade-offs within their comparison groups, while full DualSFT yields a more favorable joint-constrained trade-off than sequential hybrid baselines under matched budgets.

ClawMark: A Living-World Benchmark for Multi-Turn, Multi-Day, Multimodal Coworker Agents

Language-model agents are increasingly used as persistent coworkers that assist users across multiple working days. During such workflows, the surrounding environment may change independently of the agent: new emails arrive, calendar entries shift, knowledge-base records are updated, and evidence appears across images, scanned PDFs, audio, video, and spreadsheets. Existing benchmarks do not adequately evaluate this setting because they typically run within a single static episode and remain largely text-centric. We introduce \bench{}, a benchmark for coworker agents built around multi-turn multi-day tasks, a stateful sandboxed service environment whose state evolves between turns, and rule-based verification. The current release contains 100 tasks across 13 professional scenarios, executed against five stateful sandboxed services (filesystem, email, calendar, knowledge base, spreadsheet) and scored by 1537 deterministic Python checkers over post-execution service state; no LLM-as-judge is invoked during scoring. We benchmark seven frontier agent systems. The strongest model reaches 75.8 weighted score, but the best strict Task Success is only 20.0\%, indicating that partial progress is common while complete end-to-end workflow completion remains rare. Turn-level analysis shows that performance drops after the first exogenous environment update, highlighting adaptation to changing state as a key open challenge. We release the benchmark, evaluation harness, and construction pipeline to support reproducible coworker-agent evaluation.

Escher-Loop: Mutual Evolution by Closed-Loop Self-Referential Optimization

While recent autonomous agents demonstrate impressive capabilities, they predominantly rely on manually scripted workflows and handcrafted heuristics, inherently limiting their potential for open-ended improvement. To address this, we propose Escher-Loop, a fully closed-loop framework that operationalizes the mutual evolution of two distinct populations: Task Agents that solve concrete problems, and Optimizer Agents that recursively refine both the task agents and themselves. To sustain this self-referential evolution, we propose a dynamic benchmarking mechanism that seamlessly reuses the empirical scores of newly generated task agents as relative win-loss signals to update optimizers' scores. This mechanism leverages the evolution of task agents as an inherent signal to drive the evaluation and refinement of optimizers without additional overhead. Empirical evaluations on mathematical optimization problems demonstrate that Escher-Loop effectively pushes past the performance ceilings of static baselines, achieving the highest absolute peak performance across all evaluated tasks under matched compute. Remarkably, we observe that the optimizer agents dynamically adapt their strategies to match the shifting demands of high-performing task agents, which explains the system's continuous improvement and superior late-stage performance.

T-800: An 800 Hz Data Glove for Precise Hand Gesture Tracking

Human dexterity relies on rapid, sub-second motor adjustments, yet capturing these high-frequency dynamics remains an enduring challenge in biomechanics and robotics. Existing motion capture paradigms are compromised by a trade-off between temporal resolution and visual occlusion, failing to record the fine-grained hand motion of fast, contact-rich manipulation. Here we introduce T-800, a high-bandwidth data glove system that achieves synchronized, full-hand motion tracking at 800 Hz. By integrating a novel broadcast-based synchronization mechanism with a mechanical stress isolation architecture, our system maintains sub-frame temporal alignment across 18 distributed inertial measurement units (IMUs) during extended, vigorous movements. We demonstrate that T-800 recovers fine-grained manipulation details previously lost to temporal undersampling. Our analysis reveals that human dexterity exhibits significantly high-frequency motion energy (>100 Hz) that was fundamentally inaccessible due to the Nyquist sampling limit imposed by previous hardware constraints. To validate the system's utility for robotic manipulation, we implement a kinematic retargeting algorithm that maps T-800's high-fidelity human gestures onto dexterous robotic hand models. This demonstrates that the high-frequency motion data can be accurately translated while respecting the kinematic constraints of robotic hands, providing the rich behavioral data necessary for training robust control policies in the future.

Graph In-Context Operator Networks for Generalizable Spatiotemporal Prediction

In-context operator learning enables neural networks to infer solution operators from contextual examples without weight updates. While prior work has demonstrated the effectiveness of this paradigm in leveraging vast datasets, a systematic comparison against single-operator learning using identical training data has been absent. We address this gap through controlled experiments comparing in-context operator learning against classical operator learning (single-operator models trained without contextual examples), under the same training steps and dataset. To enable this investigation on real-world spatiotemporal systems, we propose GICON (Graph In-Context Operator Network), combining graph message passing for geometric generalization with example-aware positional encoding for cardinality generalization. Experiments on air quality prediction across two Chinese regions show that in-context operator learning outperforms classical operator learning on complex tasks, generalizing across spatial domains and scaling robustly from few training examples to 100 at inference.