Alignment, interpretability, red-teaming, bias and privacy: the research on what these systems do when they misbehave.
arXiv:2508. 17320v3 Announce Type: replace Abstract: Understanding the internal representations of large language models (LLMs) remains a central challenge for interpretability research.
arXiv:2605. 03403v2 Announce Type: replace-cross Abstract: Group Relative Policy Optimization (GRPO) has recently shown strong performance in post-training large language models and vision-language models.
arXiv:2512. 12997v2 Announce Type: replace-cross Abstract: CLIP delivers strong zero-shot classification but remains highly vulnerable to adversarial attacks.
arXiv:2606. 00947v1 Announce Type: cross Abstract: Foundation models are increasingly personalized on decentralized private data through federated learning and are now deployed at scale under growing regulatory requirements for post-market monitoring.
arXiv:2606. 02339v1 Announce Type: new Abstract: Entropy minimization (EM) is the dominant objective for test-time adaptation, yet its failure mode, model collapse, remains poorly understood.
arXiv:2606. 00572v1 Announce Type: new Abstract: Passenger count data from public transit systems reveals urban mobility patterns and is essential for planning, operation, and optimisation.
arXiv:2606. 02237v1 Announce Type: new Abstract: Distribution Matching Distillation (DMD) compresses pretrained diffusion models into efficient few-step generators by aligning their noised distributions across all scales.
arXiv:2605. 26397v2 Announce Type: replace-cross Abstract: Safety alignment reduces explicitly harmful outputs but inadvertently encodes a sanitized, neuronormative representation of marginalized communication.
arXiv:2604. 01841v2 Announce Type: replace Abstract: Clinical prediction from structured electronic health records (EHRs) is challenging due to high dimensionality, heterogeneity, class imbalance, and distribution shift.
arXiv:2606. 01160v1 Announce Type: new Abstract: Large Language Models (LLMs) are increasingly used with formal interactive theorem provers such as Lean 4.
arXiv:2606. 01028v1 Announce Type: new Abstract: Medical treatment recommendation poses several challenges to reinforcement learning (RL): patient physiology evolves in continuous time, measurements and interventions are performed at irregular intervals, and treatment effects vary substantially across individuals.
arXiv:2606. 00747v1 Announce Type: cross Abstract: For low-altitude Unmanned Aerial Vehicle (UAV) autonomy, 3D spatial understanding is not merely a perception objective, but the safety interface between human instructions and physical flight.
arXiv:2606. 00827v1 Announce Type: cross Abstract: Strategic classification (SC) investigates scenarios where agents manipulate their features to obtain favorable decisions from predictive models.
arXiv:2606. 01472v1 Announce Type: cross Abstract: High-stakes production document-generation systems require language models to be adaptive, evidence-grounded, and auditable.
arXiv:2606. 01051v1 Announce Type: new Abstract: Dynamic medical treatment requires deciding treatment intensity and intervention timing, while patient states evolve continuously and adverse events may occur between clinical interactions.
arXiv:2606. 01494v1 Announce Type: cross Abstract: Agent skills extend AI agents with reusable instructions, tools, scripts, references, and workflows, establishing a security boundary distinct from both model safety and traditional package-malware detection.
arXiv:2505. 17630v4 Announce Type: replace-cross Abstract: Circuit localization methods aim to identify the subset of model components responsible for specific behaviors in large language models, enabling detailed mechanistic analysis.
arXiv:2509. 24696v2 Announce Type: replace-cross Abstract: Personalizing large language models (LLMs) to individual user preferences is a critical step beyond generating generically helpful responses.
arXiv:2602. 02470v2 Announce Type: replace Abstract: Autoregressive large language models (LLMs) have achieved remarkable success in many complex tasks, yet they can still fail in very simple logical reasoning such as the "reversal curse" -- when trained on forward knowledge data of the form "$A \rightarrow B$" (e.
arXiv:2606. 01513v1 Announce Type: cross Abstract: High-stakes enterprise document generation, including financial dispute narratives, compliance notices, and audit summaries, demands schema correctness, policy compliance, and low-latency operation at scale.