Alignment, interpretability, red-teaming, bias and privacy: the research on what these systems do when they misbehave.
arXiv:2604. 09945v2 Announce Type: replace-cross Abstract: The rapid adoption of large vision-language models (LVLMs) in recent years has been accompanied by growing fairness concerns due to their propensity to reinforce harmful societal stereotypes.
arXiv:2606. 23370v2 Announce Type: replace-cross Abstract: Device-side Large Language Models (LLMs) have grown explosively, offering stronger privacy and higher availability than their cloud-side counterparts.
arXiv:2408. 01139v4 Announce Type: replace Abstract: Perturbation robustness evaluates the vulnerabilities of models, arising from a variety of perturbations, such as data corruptions and adversarial attacks.
arXiv:2607. 01235v1 Announce Type: cross Abstract: Understanding how Large Language Models (LLMs) make token-level decisions during code generation remains a major challenge for both researchers and practitioners.
arXiv:2607. 02510v1 Announce Type: new Abstract: Despite alignment training, LLMs remain prone to generating unsafe outputs at deployment time.
arXiv:2606. 30970v2 Announce Type: replace Abstract: Autonomous AI agents increasingly perform consequential actions on behalf of human principals, including financial transactions, external communications, and enterprise workflows.
arXiv:2605. 11020v2 Announce Type: replace-cross Abstract: Inverse reinforcement learning (IRL) is typically formulated as maximizing entropy subject to matching the distribution of expert trajectories.
arXiv:2607. 02046v1 Announce Type: new Abstract: Anomaly detection is a critical and evolving field in Machine Learning, with applications targeting different domains such as cybersecurity, finance, healthcare, manufacturing and IoT (Internet of Things) systems.
arXiv:2607. 02187v1 Announce Type: new Abstract: Distributed machine learning enables collaborative model training without centralizing data, but it also exposes learning processes to privacy leakage and malicious manipulation.
arXiv:2607. 02047v1 Announce Type: cross Abstract: Safe completion requires models to provide useful assistance without enabling harm, but this behavior is difficult to evaluate with isolated prompts.
arXiv:2607. 01365v1 Announce Type: cross Abstract: Given one or more images of a railway crossing, can we leverage visual cues that allow us to robustly estimate how safe it is?
arXiv:2607. 01586v1 Announce Type: cross Abstract: Vision-language-action models (VLAs) have recently advanced robotic manipulation, yet the effects of different robot-data pre-training paradigms remain difficult to compare because existing models often differ in architecture, data, action space, and evaluation protocol.
arXiv:2607. 01919v1 Announce Type: new Abstract: Agentic systems enhance their capabilities by invoking external tools and maintaining persistent memory.
arXiv:2604. 08169v2 Announce Type: replace Abstract: Alignment in LLMs is more brittle than commonly assumed: misalignment can be induced by adversarial prompts, benign fine-tuning, emergent misalignment, and goal misgeneralization.
arXiv:2601. 02813v3 Announce Type: replace Abstract: Aligning language models to qualitative behavioral traits, such as human-likeness, remains difficult because they are hard to define, measure, and optimize.
arXiv:2607. 02431v1 Announce Type: cross Abstract: Reinforcement learning (RL) can overcome the demonstration-coverage limitation of imitation learning (IL) by allowing robots to improve through trial-and-error interaction beyond the states observed in demonstrations.
arXiv:2607. 02426v1 Announce Type: cross Abstract: Federated learning (FL) enables collaborative model training across distributed devices without sharing raw data, making it suitable for privacy-sensitive robotic sensing applications.
arXiv:2605. 29283v2 Announce Type: replace-cross Abstract: Recent physics foundation models claim general spatiotemporal forecasting ability, yet their evaluations often collapse performance into a single average score under a fixed training distribution.
arXiv:2607. 01527v1 Announce Type: cross Abstract: Room embeddings derived from reverberant speech are often unreliable: speech content and recording degradation can alter the representation even when speaker, room, and source-receiver geometry remain unchanged, degrading downstream task performance.
arXiv:2607. 02396v1 Announce Type: new Abstract: Steering and monitoring activations in Large Language Models (LLMs) are increasingly used for both safety and interpretability.