260314 arxiv 모음
Examining Reasoning LLMs-as-Judges in Non-Verifiable LLM Post-Training
Examining Reasoning LLMs-as-Judges in Non-Verifiable LLM Post-Training
Reasoning LLMs-as-Judges, which can benefit from inference-time scaling, provide a promising path for extending the success of reasoning models to non-verifiable domains where the output correctness/quality cannot be directly checked. However, while reasoning judges have shown better performance on static evaluation benchmarks, their effectiveness in actual policy training has not been systematically examined. Therefore, we conduct a rigorous study to investigate the actual impact of non-reasoni
출처: https://arxiv.org/abs/2603.12246v1
Security Considerations for Artificial Intelligence Agents
Security Considerations for Artificial Intelligence Agents
This article, a lightly adapted version of Perplexity's response to NIST/CAISI Request for Information 2025-0035, details our observations and recommendations concerning the security of frontier AI agents. These insights are informed by Perplexity's experience operating general-purpose agentic systems used by millions of users and thousands of enterprises in both controlled and open-world environments. Agent architectures change core assumptions around code-data separation, authority boundaries,
출처: https://arxiv.org/abs/2603.12230v1
Language Model Teams as Distributed Systems
Language Model Teams as Distributed Systems
Large language models (LLMs) are growing increasingly capable, prompting recent interest in LLM teams. Yet, despite increased deployment of LLM teams at scale, we lack a principled framework for addressing key questions such as when a team is helpful, how many agents to use, how structure impacts performance -- and whether a team is better than a single agent. Rather than designing and testing these possibilities through trial-and-error, we propose using distributed systems as a principled found
출처: https://arxiv.org/abs/2603.12229v1
Sparking Scientific Creativity via LLM-Driven Interdisciplinary Inspiration
Sparking Scientific Creativity via LLM-Driven Interdisciplinary Inspiration
Despite interdisciplinary research leading to larger and longer-term impact, most work remains confined to single-domain academic silos. Recent AI-based approaches to scientific discovery show promise for interdisciplinary research, but many prioritize rapidly designing experiments and solutions, bypassing the exploratory, collaborative reasoning processes that drive creative interdisciplinary breakthroughs. As a result, prior efforts largely prioritize automating scientific discovery rather tha
출처: https://arxiv.org/abs/2603.12226v1
CLASP: Defending Hybrid Large Language Models Against Hidden State Poisoning Att
CLASP: Defending Hybrid Large Language Models Against Hidden State Poisoning Att
State space models (SSMs) like Mamba have gained significant traction as efficient alternatives to Transformers, achieving linear complexity while maintaining competitive performance. However, Hidden State Poisoning Attacks (HiSPAs), a recently discovered vulnerability that corrupts SSM memory through adversarial strings, pose a critical threat to these architectures and their hybrid variants. Framing the HiSPA mitigation task as a binary classification problem at the token level, we introduce t
출처: https://arxiv.org/abs/2603.12206v1
IndexCache: Accelerating Sparse Attention via Cross-Layer [[INDEX]] Reuse
IndexCache: Accelerating Sparse Attention via Cross-Layer Index Reuse
Long-context agentic workflows have emerged as a defining use case for large language models, making attention efficiency critical for both inference speed and serving cost. Sparse attention addresses this challenge effectively, and DeepSeek Sparse Attention (DSA) is a representative production-grade solution: a lightweight lightning indexer selects the top-k most relevant tokens per query, reducing core attention from $O(L^2)$ to $O(Lk)$. However, the indexer itself retains $O(L^2)$ complexity
출처: https://arxiv.org/abs/2603.12201v1
Long-Context Encoder Models for Polish Language Understanding
Long-Context Encoder Models for Polish Language Understanding
While decoder-only Large Language Models (LLMs) have recently dominated the NLP landscape, encoder-only architectures remain a cost-effective and parameter-efficient standard for discriminative tasks. However, classic encoders like BERT are limited by a short context window, which is insufficient for processing long documents. In this paper, we address this limitation for the Polish by introducing a high-quality Polish model capable of processing sequences of up to 8192 tokens. The model was dev
출처: https://arxiv.org/abs/2603.12191v1
BehaviorVLM: Unified Finetuning-Free Behavioral Understanding with Vision-Langua
BehaviorVLM: Unified Finetuning-Free Behavioral Understanding with Vision-Langua
Understanding freely moving animal behavior is central to neuroscience, where pose estimation and behavioral understanding form the foundation for linking neural activity to natural actions. Yet both tasks still depend heavily on human annotation or unstable unsupervised pipelines, limiting scalability and reproducibility. We present BehaviorVLM, a unified vision-language framework for pose estimation and behavioral understanding that requires no task-specific finetuning and minimal human labeli
출처: https://arxiv.org/abs/2603.12176v1
Deep Incentive Design with Differentiable Equilibrium Blocks
Deep Incentive Design with Differentiable Equilibrium Blocks
Automated design of multi-agent interactions with desirable equilibrium outcomes is inherently difficult due to the computational hardness, non-uniqueness, and instability of the resulting equilibria. In this work, we propose the use of game-agnostic differentiable equilibrium blocks (DEBs) as modules in a novel, differentiable framework to address a wide variety of incentive design problems from economics and computer science. We call this framework deep incentive design (DID). To validate our
출처: https://arxiv.org/abs/2603.07705v2
QAQ: Bidirectional Semantic Coherence for Selecting High-Quality Synthetic Code
QAQ: Bidirectional Semantic Coherence for Selecting High-Quality Synthetic Code
Synthetic data has become essential for training code generation models, yet it introduces significant noise and hallucinations that are difficult to detect with current metrics. Existing data selection methods like Instruction-Following Difficulty (IFD) typically assess how hard a model generates an answer given a query ($A|Q$). However, this metric is ambiguous on noisy synthetic data, where low probability can distinguish between intrinsic task complexity and model-generated hallucinations. H
출처: https://arxiv.org/abs/2603.12165v1
NeuralOS: Towards Simulating Operating Systems via Neural Generative Models
NeuralOS: Towards Simulating Operating Systems via Neural Generative Models
We introduce NeuralOS, a neural framework that simulates graphical user interfaces (GUIs) of operating systems by directly predicting screen frames in response to user inputs such as mouse movements, clicks, and keyboard events. NeuralOS combines a recurrent neural network (RNN), which tracks computer state, with a diffusion-based neural renderer that generates screen images. The model is trained on a dataset of Ubuntu XFCE recordings, which include both randomly generated interactions and reali
출처: https://arxiv.org/abs/2507.08800v2
Separable neural architectures as a primitive for unified predictive and generat
Separable neural architectures as a primitive for unified predictive and generat
Intelligent systems across physics, language and perception often exhibit factorisable structure, yet are typically modelled by monolithic neural architectures that do not explicitly exploit this structure. The separable neural architecture (SNA) addresses this by formalising a representational class that unifies additive, quadratic and tensor-decomposed neural models. By constraining interaction order and tensor rank, SNAs impose a structural inductive bias that factorises high-dimensional mapp
출처: https://arxiv.org/abs/2603.12244v1
Strategic Navigation or Stochastic Search? How Agents and Humans Reason Over Doc
Strategic Navigation or Stochastic Search? How Agents and Humans Reason Over Doc
Multimodal agents offer a promising path to automating complex document-intensive workflows. Yet, a critical question remains: do these agents demonstrate genuine strategic reasoning, or merely stochastic trial-and-error search? To address this, we introduce MADQA, a benchmark of 2,250 human-authored questions grounded in 800 heterogeneous PDF documents. Guided by Classical Test Theory, we design it to maximize discriminative power across varying levels of agentic abilities. To evaluate agentic
출처: https://arxiv.org/abs/2603.12180v1
GlyphBanana: Advancing Precise Text Rendering Through Agentic Workflows
GlyphBanana: Advancing Precise Text Rendering Through Agentic Workflows
Despite recent advances in generative models driving significant progress in text rendering, accurately generating complex text and mathematical formulas remains a formidable challenge. This difficulty primarily stems from the limited instruction-following capabilities of current models when encountering out-of-distribution prompts. To address this, we introduce GlyphBanana, alongside a corresponding benchmark specifically designed for rendering complex characters and formulas. GlyphBanana emplo
출처: https://arxiv.org/abs/2603.12155v1
Automatic Generation of High-Performance RL Environments
Automatic Generation of High-Performance RL Environments
Translating complex reinforcement learning (RL) environments into high-performance implementations has traditionally required months of specialized engineering. We present a reusable recipe - a generic prompt template, hierarchical verification, and iterative agent-assisted repair - that produces semantically equivalent high-performance environments for <$10 in compute cost. We demonstrate three distinct workflows across five environments. Direct translation (no prior performance implementation
출처: https://arxiv.org/abs/2603.12145v1
Increasing intelligence in AI agents can worsen collective outcomes
Increasing intelligence in AI agents can worsen collective outcomes
When resources are scarce, will a population of AI agents coordinate in harmony, or descend into tribal chaos? Diverse decision-making AI from different developers is entering everyday devices -- from phones and medical devices to battlefield drones and cars -- and these AI agents typically compete for finite shared resources such as charging slots, relay bandwidth, and traffic priority. Yet their collective dynamics and hence risks to users and society are poorly understood. Here we study AI-ag
출처: https://arxiv.org/abs/2603.12129v1
Matching Features, Not Tokens: Energy-Based Fine-Tuning of Language Models
Matching Features, Not Tokens: Energy-Based Fine-Tuning of Language Models
Cross-entropy (CE) training provides dense and scalable supervision for language models, but it optimizes next-token prediction under teacher forcing rather than sequence-level behavior under model rollouts. We introduce a feature-matching objective for language-model fine-tuning that targets sequence-level statistics of the completion distribution, providing dense semantic feedback without requiring a task-specific verifier or preference model. To optimize this objective efficiently, we propose
출처: https://arxiv.org/abs/2603.12248v1
BiGain: Unified Token Compression for Joint Generation and Classification
BiGain: Unified Token Compression for Joint Generation and Classification
Acceleration methods for diffusion models (e.g., token merging or downsampling) typically optimize synthesis quality under reduced compute, yet often ignore discriminative capacity. We revisit token compression with a joint objective and present BiGain, a training-free, plug-and-play framework that preserves generation quality while improving classification in accelerated diffusion models. Our key insight is frequency separation: mapping feature-space signals into a frequency-aware representatio
출처: https://arxiv.org/abs/2603.12240v1
STAMP: Selective Task-Aware Mechanism for Text Privacy
STAMP: Selective Task-Aware Mechanism for Text Privacy
We present STAMP (Selective Task-Aware Mechanism for Text Privacy), a new framework for task-aware text privatization that achieves an improved privacy-utility trade-off. STAMP selectively allocates privacy budgets across tokens by jointly considering (i) each token's importance to the downstream task (as measured via a task- or query-specific representation), and (ii) its privacy sensitivity (e.g., names, dates, identifiers). This token-level partitioning enables fine-grained, group-wise contro
출처: https://arxiv.org/abs/2603.12237v1
Interpreting Contrastive Embeddings in Specific Domains with Fuzzy Rules
Interpreting Contrastive Embeddings in Specific Domains with Fuzzy Rules
Free-style text is still one of the common ways in which data is registered in real environments, like legal procedures and medical records. Because of that, there have been significant efforts in the area of natural language processing to convert these texts into a structured format, which standard machine learning methods can then exploit. One of the most popular methods to embed text into a vectorial representation is the Contrastive Language-Image Pre-training model (CLIP), which was trained
출처: https://arxiv.org/abs/2603.12227v1
HOG-Diff: Higher-Order Guided Diffusion for Graph Generation
HOG-Diff: Higher-Order Guided Diffusion for Graph Generation
Graph generation is a critical yet challenging task, as empirical analyses require a deep understanding of complex, non-Euclidean structures. Diffusion models have recently made significant advances in graph generation, but these models are typically adapted from image generation frameworks and overlook inherent higher-order topology, limiting their ability to capture graph topology. In this work, we propose Higher-order Guided Diffusion (HOG-Diff), a principled framework that progressively gene
출처: https://arxiv.org/abs/2502.04308v3
Expert Selections In MoE Models Reveal (Almost) As Much As Text
Expert Selections In MoE Models Reveal (Almost) As Much As Text
We present a text-reconstruction attack on mixture-of-experts (MoE) language models that recovers tokens from expert selections alone. In MoE models, each token is routed to a subset of expert subnetworks; we show these routing decisions leak substantially more information than previously understood. Prior work using logistic regression achieves limited reconstruction; we show that a 3-layer MLP improves this to 63.1% top-1 accuracy, and that a transformer-based sequence decoder recovers 91.2% o
출처: https://arxiv.org/abs/2602.04105v2
The Latent Color Subspace: Emergent Order in High-Dimensional Chaos
The Latent Color Subspace: Emergent Order in High-Dimensional Chaos
Text-to-image generation models have advanced rapidly, yet achieving fine-grained control over generated images remains difficult, largely due to limited understanding of how semantic information is encoded. We develop an interpretation of the color representation in the Variational Autoencoder latent space of FLUX.1 [Dev], revealing a structure reflecting Hue, Saturation, and Lightness. We verify our Latent Color Subspace (LCS) interpretation by demonstrating that it can both predict and explic
출처: https://arxiv.org/abs/2603.12261v1
SciMDR: Benchmarking and Advancing Scientific Multimodal Document Reasoning
SciMDR: Benchmarking and Advancing Scientific Multimodal Document Reasoning
Constructing scientific multimodal document reasoning datasets for foundation model training involves an inherent trade-off among scale, faithfulness, and realism. To address this challenge, we introduce the synthesize-and-reground framework, a two-stage pipeline comprising: (1) Claim-Centric QA Synthesis, which generates faithful, isolated QA pairs and reasoning on focused segments, and (2) Document-Scale Regrounding, which programmatically re-embeds these pairs into full-document tasks to ensu
출처: https://arxiv.org/abs/2603.12249v1
Incremental Neural Network Verification via Learned Conflicts
Incremental Neural Network Verification via Learned Conflicts
Neural network verification is often used as a core component within larger analysis procedures, which generate sequences of closely related verification queries over the same network. In existing neural network verifiers, each query is typically solved independently, and information learned during previous runs is discarded, leading to repeated exploration of the same infeasible regions of the search space. In this work, we aim to expedite verification by reducing this redundancy. We propose an
출처: https://arxiv.org/abs/2603.12232v1
Temporal Straightening for Latent Planning
Temporal Straightening for Latent Planning
Learning good representations is essential for latent planning with world models. While pretrained visual encoders produce strong semantic visual features, they are not tailored to planning and contain information irrelevant -- or even detrimental -- to planning. Inspired by the perceptual straightening hypothesis in human visual processing, we introduce temporal straightening to improve representation learning for latent planning. Using a curvature regularizer that encourages locally straighten
출처: https://arxiv.org/abs/2603.12231v1
Neural Thickets: Diverse Task Experts Are Dense Around Pretrained Weights
Neural Thickets: Diverse Task Experts Are Dense Around Pretrained Weights
Pretraining produces a learned parameter vector that is typically treated as a starting point for further iterative adaptation. In this work, we instead view the outcome of pretraining as a distribution over parameter vectors, whose support already contains task-specific experts. We show that in small models such expert solutions occupy a negligible fraction of the volume of this distribution, making their discovery reliant on structured optimization methods such as gradient descent. In contrast
출처: https://arxiv.org/abs/2603.12228v1
HiAP: A Multi-Granular Stochastic Auto-Pruning Framework for Vision Transformers
HiAP: A Multi-Granular Stochastic Auto-Pruning Framework for Vision Transformers
Vision Transformers require significant computational resources and memory bandwidth, severely limiting their deployment on edge devices. While recent structured pruning methods successfully reduce theoretical FLOPs, they typically operate at a single structural granularity and rely on complex, multi-stage pipelines with post-hoc thresholding to satisfy sparsity budgets. In this paper, we propose Hierarchical Auto-Pruning (HiAP), a continuous relaxation framework that discovers optimal sub-netwo
출처: https://arxiv.org/abs/2603.12222v1
Spatial-TTT: Streaming Visual-based Spatial Intelligence with Test-Time Training
Spatial-TTT: Streaming Visual-based Spatial Intelligence with Test-Time Training
Humans perceive and understand real-world spaces through a stream of visual observations. Therefore, the ability to streamingly maintain and update spatial evidence from potentially unbounded video streams is essential for spatial intelligence. The core challenge is not simply longer context windows but how spatial information is selected, organized, and retained over time. In this paper, we propose Spatial-TTT towards streaming visual-based spatial intelligence with test-time training (TTT), wh
출처: https://arxiv.org/abs/2603.12255v1
Quantifying Aleatoric Uncertainty of the Treatment Effect: A Novel Orthogonal Le
Quantifying Aleatoric Uncertainty of the Treatment Effect: A Novel Orthogonal Le
Estimating causal quantities from observational data is crucial for understanding the safety and effectiveness of medical treatments. However, to make reliable inferences, medical practitioners require not only estimating averaged causal quantities, such as the conditional average treatment effect, but also understanding the randomness of the treatment effect as a random variable. This randomness is referred to as aleatoric uncertainty and is necessary for understanding the probability of benefi
출처: https://arxiv.org/abs/2411.03387v3
관련 노트
- [[260324_arxiv]]
- [[260323_arxiv]]
- [[INDEX]]
- [[260318_arxiv]] — 키워드 유사
- [[260317_arxiv]] — 키워드 유사
- [[260316_x]] — 키워드 유사