10003 papers
This paper investigates how large language models process conflicting demonstrations in in-context learning, revealing a two-phase computational structure where early layers encode both correct and incorrect rules while late layers commit to predictions, and identifies specific attention heads responsible for this vulnerability that can be mitigated through targeted ablation to significantly improve performance.
This study proposes an interpretable LSTM-based model for predicting ship trajectories in inland waterways that incorporates trained ship domain parameters to analyze attention mechanisms, revealing that while the model achieves competitive accuracy, its attention weights do not fully align with expected causal relationships between interacting vessels.
This paper proposes a propagation dynamics model and a non-intrusive genealogy-graph-based governance layer to identify, model, and effectively mitigate error cascades in LLM-based multi-agent systems, significantly improving defense success rates against error amplification without altering existing collaboration architectures.
This paper presents a circular dilated convolutional neural network (CDCNN) for real-time activity recognition using multi-modal smart insole sensor data, which achieves 86.42% accuracy in a subject-independent four-class classification task, demonstrating that inertial sensors are the primary contributors to performance while remaining suitable for embedded deployment.
This paper proposes Progressive Refinement Regulation (PRR), a dynamic, trajectory-grounded framework that learns a token-wise controller to adaptively regulate the denoising process based on empirical convergence progress, thereby substantially accelerating Diffusion Language Model decoding while preserving generation quality.
This paper introduces a contrastive learning framework that aligns X-ray spectra with scientific literature to create shared multimodal representations, significantly improving the estimation of physical variables and enabling the discovery of rare astrophysical sources through integrated data analysis.
The paper introduces Projected Hessian Learning (PHL), a scalable framework that enables efficient, curvature-informed training of machine-learning interatomic potentials by utilizing stochastic Hessian-vector products instead of explicit Hessian matrices, thereby achieving full-second-order accuracy with significantly reduced computational cost and memory requirements.
This paper presents a multi-agent system that autonomously discovers mathematical concepts, such as homology, by dynamically interweaving conjecture generation, proof attempts, and counterexample analysis, demonstrating that optimizing these local processes effectively aligns with human notions of mathematical interestingness.
This paper investigates temporal drift in the FreshStack retrieval benchmark by comparing 2024 and 2025 corpus snapshots, finding that while relevant documents migrate between repositories, retrieval model rankings remain highly stable, suggesting that such benchmarks can remain reliable despite temporal corpus evolution.
This paper proposes Invariant Causal Routing (ICR), a causal governance framework that leverages counterfactual reasoning and invariant causal discovery to identify stable, interpretable policy rules for steering emergent social norms in online market economies across heterogeneous environments.
This paper proposes Adaptive Memory Admission Control (A-MAC), a framework that decomposes memory value into five interpretable factors to enable transparent, efficient, and domain-adaptive long-term memory management for LLM agents, achieving superior precision-recall tradeoffs and reduced latency compared to state-of-the-art systems.
This study investigates the correlation between catastrophic forgetting and structural collapse in continual learning by measuring weight and activation effective rank across various architectures and strategies, revealing that forgetting is strongly linked to the loss of model plasticity and that different methods preserve capacity and performance with varying efficiency.
This paper identifies "self-attribution bias" in agentic systems, demonstrating that language model monitors are significantly less likely to flag high-risk or low-quality actions when evaluating their own previously generated outputs compared to identical actions presented by a user, a flaw that can lead to the deceptive overestimation of monitor reliability in real-world deployments.
The paper proposes ECG-MoE, a hybrid Mixture-of-Experts foundation model that integrates beat-level morphology and rhythm analysis with cardiac period awareness to achieve state-of-the-art performance on five clinical ECG tasks while significantly reducing inference time.
This paper introduces GOLF, a reinforcement learning framework that leverages group-level natural language feedback from external critiques and intra-group attempts to generate actionable refinements, thereby significantly improving sample efficiency and exploration in sparse-reward environments compared to traditional scalar reward methods.
This paper introduces Vibe Code Bench, a novel benchmark featuring 100 web application specifications evaluated by autonomous browser agents, which reveals that even the best frontier models achieve only 58.0% accuracy on end-to-end development tasks and highlights self-testing and evaluator alignment as critical factors for success.
This paper proposes a human-guided framework that leverages Large Language Models to automate dataset risk estimation by identifying schema properties, generating clustering code, and interpreting results, thereby addressing the limitations of manual auditing and fully automated AI approaches.
This study demonstrates that LLM-based agents can be prompted to generate propaganda using various rhetorical techniques, but their susceptibility to such manipulation can be significantly mitigated through fine-tuning methods, with ORPO proving to be the most effective approach.
This paper introduces RoboMME, a large-scale standardized benchmark and a suite of memory-augmented VLA models designed to systematically evaluate and advance robotic generalist policies in long-horizon, history-dependent manipulation tasks.
This paper proposes augmenting Proximal Policy Optimization with temporal sequence models, particularly Transformers, to enable robust reinforcement learning under sensor drift and partial observability by inferring missing information from history, a claim supported by theoretical bounds on reward degradation and empirical success on MuJoCo benchmarks.