MedMT-Bench: Can LLMs Memorize and Understand Long Multi-Turn Conversations in Medical Scenarios?
arXiv:2603.23519v1 Announce Type: new Abstract: Large Language Models (LLMs) have demonstrated impressive capabilities across various specialist domains and have been integrated into high-stakes areas such as medicine. However, as existing medical-related benchmarks rarely stress-test the long-context memory, interference robustness, and...
Internal Safety Collapse in Frontier Large Language Models
arXiv:2603.23509v1 Announce Type: new Abstract: This work identifies a critical failure mode in frontier large language models (LLMs), which we term Internal Safety Collapse (ISC): under certain task conditions, models enter a state in which they continuously generate harmful content...
Navigating the Concept Space of Language Models
arXiv:2603.23524v1 Announce Type: new Abstract: Sparse autoencoders (SAEs) trained on large language model activations output thousands of features that enable mapping to human-interpretable concepts. The current practice for analyzing these features primarily relies on inspecting top-activating examples, manually browsing individual...
Fast and Faithful: Real-Time Verification for Long-Document Retrieval-Augmented Generation Systems
arXiv:2603.23508v1 Announce Type: new Abstract: Retrieval-augmented generation (RAG) is increasingly deployed in enterprise search and document-centric assistants, where responses must be grounded in long and complex source materials. In practice, verifying that generated answers faithfully reflect retrieved documents is difficult:...
Leveraging Computerized Adaptive Testing for Cost-effective Evaluation of Large Language Models in Medical Benchmarking
arXiv:2603.23506v1 Announce Type: new Abstract: The rapid proliferation of large language models (LLMs) in healthcare creates an urgent need for scalable and psychometrically sound evaluation methods. Conventional static benchmarks are costly to administer repeatedly, vulnerable to data contamination, and lack...
Revisiting Real-Time Digging-In Effects: No Evidence from NP/Z Garden-Paths
arXiv:2603.23624v1 Announce Type: new Abstract: Digging-in effects, where disambiguation difficulty increases with longer ambiguous regions, have been cited as evidence for self-organized sentence processing, in which structural commitments strengthen over time. In contrast, surprisal theory predicts no such effect unless...
Probing Ethical Framework Representations in Large Language Models: Structure, Entanglement, and Methodological Challenges
arXiv:2603.23659v1 Announce Type: new Abstract: When large language models make ethical judgments, do their internal representations distinguish between normative frameworks, or collapse ethics into a single acceptability dimension? We probe hidden representations across five ethical frameworks (deontology, utilitarianism, virtue, justice,...
The Diminishing Returns of Early-Exit Decoding in Modern LLMs
arXiv:2603.23701v1 Announce Type: new Abstract: In Large Language Model (LLM) inference, early-exit refers to stopping computation at an intermediate layer once the prediction is sufficiently confident, thereby reducing latency and cost. However, recent LLMs adopt improved pretraining recipes and architectures...
Infrequent Child-Directed Speech Is Bursty and May Draw Infant Vocalizations
arXiv:2603.23797v1 Announce Type: new Abstract: Children in many parts of the world hear relatively little speech directed to them, yet still reach major language development milestones. What differs about the speech input that infants learn from when directed input is...
Perturbation: A simple and efficient adversarial tracer for representation learning in language models
arXiv:2603.23821v1 Announce Type: new Abstract: Linguistic representation learning in deep neural language models (LMs) has been studied for decades, for both practical and theoretical reasons. However, finding representations in LMs remains an unsolved problem, in part due to a dilemma...
Language Model Planners do not Scale, but do Formalizers?
arXiv:2603.23844v1 Announce Type: new Abstract: Recent work shows overwhelming evidence that LLMs, even those trained to scale their reasoning trace, perform unsatisfactorily when solving planning problems too complex. Whether the same conclusion holds for LLM formalizers that generate solver-oriented programs...
BeliefShift: Benchmarking Temporal Belief Consistency and Opinion Drift in LLM Agents
arXiv:2603.23848v1 Announce Type: new Abstract: LLMs are increasingly used as long-running conversational agents, yet every major benchmark evaluating their memory treats user information as static facts to be stored and retrieved. That's the wrong model. People change their minds, and...
OmniACBench: A Benchmark for Evaluating Context-Grounded Acoustic Control in Omni-Modal Models
arXiv:2603.23938v1 Announce Type: new Abstract: Most testbeds for omni-modal models assess multimodal understanding via textual outputs, leaving it unclear whether these models can properly speak their answers. To study this, we introduce OmniACBench, a benchmark for evaluating context-grounded acoustic control...
Argument Mining as a Text-to-Text Generation Task
arXiv:2603.23949v1 Announce Type: new Abstract: Argument Mining(AM) aims to uncover the argumentative structures within a text. Previous methods require several subtasks, such as span identification, component classification, and relation classification. Consequently, these methods need rule-based postprocessing to derive argumentative structures...
From AI Assistant to AI Scientist: Autonomous Discovery of LLM-RL Algorithms with LLM Agents
arXiv:2603.23951v1 Announce Type: new Abstract: Discovering improved policy optimization algorithms for language models remains a costly manual process requiring repeated mechanism-level modification and validation. Unlike simple combinatorial code search, this problem requires searching over algorithmic mechanisms tightly coupled with training...
Thinking with Tables: Enhancing Multi-Modal Tabular Understanding via Neuro-Symbolic Reasoning
arXiv:2603.24004v1 Announce Type: new Abstract: Multimodal Large Language Models (MLLMs) have demonstrated remarkable reasoning capabilities across modalities such as images and text. However, tabular data, despite being a critical real-world modality, remains relatively underexplored in multimodal learning. In this paper,...
Implicit Turn-Wise Policy Optimization for Proactive User-LLM Interaction
arXiv:2603.23550v1 Announce Type: new Abstract: Multi-turn human-AI collaboration is fundamental to deploying interactive services such as adaptive tutoring, conversational recommendation, and professional consultation. However, optimizing these interactions via reinforcement learning is hindered by the sparsity of verifiable intermediate rewards and...
Upper Entropy for 2-Monotone Lower Probabilities
arXiv:2603.23558v1 Announce Type: new Abstract: Uncertainty quantification is a key aspect in many tasks such as model selection/regularization, or quantifying prediction uncertainties to perform active learning or OOD detection. Within credal approaches that consider modeling uncertainty as probability sets, upper...
Synthetic Mixed Training: Scaling Parametric Knowledge Acquisition Beyond RAG
arXiv:2603.23562v1 Announce Type: new Abstract: Synthetic data augmentation helps language models learn new knowledge in data-constrained domains. However, naively scaling existing synthetic data methods by training on more synthetic tokens or using stronger generators yields diminishing returns below the performance...
Safe Reinforcement Learning with Preference-based Constraint Inference
arXiv:2603.23565v1 Announce Type: new Abstract: Safe reinforcement learning (RL) is a standard paradigm for safety-critical decision making. However, real-world safety constraints can be complex, subjective, and even hard to explicitly specify. Existing works on constraint inference rely on restrictive assumptions...
AscendOptimizer: Episodic Agent for Ascend NPU Operator Optimization
arXiv:2603.23566v1 Announce Type: new Abstract: AscendC (Ascend C) operator optimization on Huawei Ascend neural processing units (NPUs) faces a two-fold knowledge bottleneck: unlike the CUDA ecosystem, there are few public reference implementations to learn from, and performance hinges on a...
StateLinFormer: Stateful Training Enhancing Long-term Memory in Navigation
arXiv:2603.23571v1 Announce Type: new Abstract: Effective navigation intelligence relies on long-term memory to support both immediate generalization and sustained adaptation. However, existing approaches face a dilemma: modular systems rely on explicit mapping but lack flexibility, while Transformer-based end-to-end models are...
Dual-Criterion Curriculum Learning: Application to Temporal Data
arXiv:2603.23573v1 Announce Type: new Abstract: Curriculum Learning (CL) is a meta-learning paradigm that trains a model by feeding the data instances incrementally according to a schedule, which is based on difficulty progression. Defining meaningful difficulty assessment measures is crucial and...
PoiCGAN: A Targeted Poisoning Based on Feature-Label Joint Perturbation in Federated Learning
arXiv:2603.23574v1 Announce Type: new Abstract: Federated Learning (FL), as a popular distributed learning paradigm, has shown outstanding performance in improving computational efficiency and protecting data privacy, and is widely applied in industrial image classification. However, due to its distributed nature,...
The Geometric Price of Discrete Logic: Context-driven Manifold Dynamics of Number Representations
arXiv:2603.23577v1 Announce Type: new Abstract: Large language models (LLMs) generalize smoothly across continuous semantic spaces, yet strict logical reasoning demands the formation of discrete decision boundaries. Prevailing theories relying on linear isometric projections fail to resolve this fundamental tension. In...
Residual Attention Physics-Informed Neural Networks for Robust Multiphysics Simulation of Steady-State Electrothermal Energy Systems
arXiv:2603.23578v1 Announce Type: new Abstract: Efficient thermal management and precise field prediction are critical for the design of advanced energy systems, including electrohydrodynamic transport, microfluidic energy harvesters, and electrically driven thermal regulators. However, the steady-state simulation of these electrothermal coupled...
MetaKube: An Experience-Aware LLM Framework for Kubernetes Failure Diagnosis
arXiv:2603.23580v1 Announce Type: new Abstract: Existing LLM-based Kubernetes diagnostic systems cannot learn from operational experience, operating on static knowledge bases without improving from past resolutions. We present MetaKube, an experience-aware LLM framework through three synergistic innovations: (1) an Episodic Pattern...
AI Generalisation Gap In Comorbid Sleep Disorder Staging
arXiv:2603.23582v1 Announce Type: new Abstract: Accurate sleep staging is essential for diagnosing OSA and hypopnea in stroke patients. Although PSG is reliable, it is costly, labor-intensive, and manually scored. While deep learning enables automated EEG-based sleep staging in healthy subjects,...
LineMVGNN: Anti-Money Laundering with Line-Graph-Assisted Multi-View Graph Neural Networks
arXiv:2603.23584v1 Announce Type: new Abstract: Anti-money laundering (AML) systems are important for protecting the global economy. However, conventional rule-based methods rely on domain knowledge, leading to suboptimal accuracy and a lack of scalability. Graph neural networks (GNNs) for digraphs (directed...
Steering Code LLMs with Activation Directions for Language and Library Control
arXiv:2603.23629v1 Announce Type: new Abstract: Code LLMs often default to particular programming languages and libraries under neutral prompts. We investigate whether these preferences are encoded as approximately linear directions in activation space that can be manipulated at inference time. Using...