Model releases, architecture work and prompting research on large language models — from frontier-lab announcements to the arXiv papers behind them.
An Intraoperative Hypotension (IOH) event is a frequent complication during administration of general anaesthesia with serious downstream consequences, yet clinical management remains reactive and not...
Electroencephalography (EEG) provides a non-invasive window into dynamic brain activity, yet modeling long-horizon EEG sequences remains challenging due to their high temporal complexity, substantial...
As Large Language Model (LLM) agents are applied in continuously interactive environments, driving the evolution of their own capabilities becomes a core problem for achieving long-term autonomy. Curr...
Instance segmentation is a fundamental computer vision task with diverse real-world applications. Recently, prompt-driven foundation models have shown promising generalization. However, automated prom...
Visual segmentation systems encounter objects outside their training distribution during real-world deployment, hindering reliable autonomous systems that depend on scene parsing in the perception sta...
Looped Language Models (LoopLMs) perform "latent reasoning" by recursively refining internal latent representations with shared weights, offering a more effective alternative to explicit verbal reason...
Simulation success is not equivalent to structural correctness for LLM-generated circuits. We define and measure four evaluation levels -- schema validity, topological validity, backend executability,...
Judge-specific sensitivity is useful for aggregating pairwise LLM evaluations, but its interpretation depends on which systematic presentation effects the ranking model includes. We introduce OSCAR, a...
The paper introduces an uncertainty‑driven training framework for 3D CT lung nodule classification that uses validation‑based uncertainty estimates to reweight the loss, aiming to improve predictive performance and probability calibration. Two uncertainty quantification methods—Monte Carlo Dropout and Evidential Deep Learning—are evaluated across multiple backbone architectures (ResNet, DenseNet, EfficientNet, ViT, Swin) on the LIDC‑IDRI and NoduleMNIST3D datasets. The approach yields comparable classification accuracy to conventional training while substantially reducing expected calibration error, especially on convolutional backbones, and shows that simple temperature scaling can also achieve strong calibration.
Large language models are increasingly used as automated reviewers in scientific evaluation, creating a recursive feedback loop where later reviewers learn from earlier model-generated judgments. A study using Llama 3.1 8B fine‑tuned on ICLR reviews shows that incorporating synthetic reviews compresses rating distributions and reduces semantic diversity, a phenomenon termed scientific‑judgment collapse. To counter this, the authors introduce TrustReviewer, an open‑source LLM system that curates training data and applies paired activation steering at test time to preserve judgment diversity and improve recommendation alignment.
The paper introduces Jarvis, an offline, edge‑deployable voice assistant designed for autonomous racecars. It combines speech recognition, synthesis, and a lightweight text‑to‑command classifier fine‑tuned from the Mistral 7B model to provide high‑level behavioral commands. Experiments show 97.63 % intent recognition accuracy with an average latency of 1.39 s, outperforming larger online‑hosted models and enabling quick response times for time‑critical driving tasks.
CaLR: Causal Latent Revision for Robust Diffusion Reasoning proposes a framework that combines autoregressive and diffusion language models by reformulating reasoning as constrained latent optimization. The method uses a causal topology matrix from an expert model and implicit differentiation to perform gradient-guided thought revision, enforcing logical consistency and allowing dynamic self-correction of intermediate steps during parallel generation. Empirical results show CaLR achieves state‑of‑the‑art performance on complex benchmarks, surpassing strong autoregressive baselines and demonstrating superior robustness in constrained tasks such as Sudoku.
The paper introduces a 12,000-instance dataset of two-option moral dilemmas covering three pairwise value conflicts—Honesty vs. Justice, Justice vs. Autonomy, and Autonomy vs. Honesty—translated into Hindi, Arabic, Spanish, and Chinese to test cross‑lingual behavior. Benchmarking on GPT‑5‑mini shows a consistent preference for Honesty over Autonomy across all languages when no policy is provided, while Llama‑3.2‑1/3B models exhibit a strong first‑option bias that is largely eliminated by plain fine‑tuning or Direct Preference Optimization, raising accuracy above 98%. The authors propose a task vector transfer method that orthogonalizes value preference vectors with respect to general instruction‑following vectors, effectively isolating specific value directions and enabling task arithmetic to flip a model’s stance.
The paper examines how conversational AI, specifically ChatGPT, displays aspects of cooperative dialogue such as morality, politeness, and alignment compared to human-human conversations. Using over 26,000 multi‑turn dialogues and mixed‑effects modeling, the authors find that AI mimics the surface features of cooperation—like warmth and hedging—yet lacks the underlying social architecture that drives mutual adaptation. Key findings include a dissociation between AI’s moral output and human negotiation, a decline in linguistic convergence, and a reversal of typical human accommodation mechanisms when interacting with AI.
The paper introduces a lightweight pre‑encoder gate for Transformer‑based time‑series forecasters, which assigns sigmoid scores to covariate representations before they enter the encoder. The gate is evaluated as a plug‑in for models such as TimeXer, iTransformer, and PatchTST on datasets including ETTm1, ETTm2, Traffic, Energy, and ILI, showing competitive performance and the ability to regulate covariate admission via a usage penalty. Experiments also explore gate placement, initialization, and feature importance using VIF‑informed permutation diagnostics.
The paper investigates why large language models (LLMs) struggle in strategic decision-making under incomplete information. It identifies two key gaps: an observation‑belief gap where LLMs’ internal representations of game states are accurate but brittle, and a belief‑action gap where converting these internal beliefs into actions is weak, leading to suboptimal payoffs. Experiments with Llama 3.1, Qwen3, and gpt‑oss confirm that acting optimally on decoded beliefs would improve outcomes in most games, highlighting a bottleneck in belief‑to‑action conversion.
The paper introduces a benchmarking framework that evaluates open-weight Vision‑Language Models (VLMs) for face recognition by treating explanation quality as a core metric. It defines two key criteria for explanations—relevance, meaning reliance on identity‑stable facial features, and faithfulness, meaning alignment with the visible image content without hallucinations. Using this framework, the authors benchmark several VLM families, jointly assessing face verification accuracy and explanation quality, and find that current models still exhibit shortcomings in their explanations, underscoring the importance of explanation metrics for a complete performance assessment.
OpenMAS-GCom is a diagnostic benchmark designed to isolate the impact of communication structures, role assignments, and information flows in graph‑enhanced multi‑agent systems (G‑MAS). It evaluates systems by systematically modifying one component—such as rewiring communication edges, removing specialist or critic agents, or corrupting intermediate messages—while keeping tasks, models, prompts, and budget limits constant. The benchmark tests 17 configurations across 29 datasets in six domains, including 400 new G‑MAS‑Complex tasks that require agents to combine and reconcile information from multiple documents.
The paper introduces the Intersection Euler Characteristic Profile, a topological metric for measuring class overlap in neural representations, and provides exact permutation and sign‑flip tests to assess disentanglement across layers. Using this statistic, the authors analyze 111 networks and 52,650 measurements, finding that disentanglement is depth‑graded, occurs early, and is influenced by training choices such as augmentation and weight decay. The study also demonstrates that the unnormalized mass of the profile predicts test accuracy, while the dimensionless quotient does not outperform simple linear probes.
The paper introduces HeadEntropy, a training‑free method that predicts the correctness of large language model (LLM) answers by measuring how stable each attention head’s pattern is to further gradient updates. By linking the trace of the softmax Jacobian to 2‑Renyi entropy, the authors show that attention spread correlates with gradient stability, enabling accurate hallucination detection without reference annotations. Across five instruction‑tuned LLMs and five diverse datasets—including medicine, multi‑hop reasoning, and mathematics—HeadEntropy achieves a 0.736 AUROC, outperforming other training‑free baselines and matching hidden‑state probes while incurring less than 1% of inference cost.