Alignment, interpretability, red-teaming, bias and privacy: the research on what these systems do when they misbehave.
arXiv:2508. 04225v4 Announce Type: replace-cross Abstract: Behavior Regularized Policy Optimization (BRPO) leverages asymmetric divergence regularization to mitigate distribution shift in offline reinforcement learning.
arXiv:2607. 17003v1 Announce Type: cross Abstract: Imitation learning enables learning a policy in an unknown environment with a latent reward signal using expert demonstrations, but it struggles when the imitator's and expert's observations are mismatched and unobserved confounders are present in expert demonstrations.
arXiv:2607. 17947v1 Announce Type: new Abstract: Existing AI measurement frameworks quantify cognitive capability, task automation, or catastrophic risk, but none measure autonomous agency: the extent to which a system behaves in a self-directed way.
arXiv:2607. 18027v1 Announce Type: new Abstract: Scaled dot product attention conflates directional alignment and vector magnitude, limiting its effectiveness as a similarity metric in Transformer models.
arXiv:2511. 02152v2 Announce Type: replace Abstract: Time series data is one of the most popular data modalities in critical domains such as industry and medicine.
arXiv:2603. 18046v2 Announce Type: replace-cross Abstract: We present NanoZK, a zero-knowledge proof system for verifiable LLM inference: clients and third-party auditors check that a provider executed the advertised model on a committed input without learning weights or activations.
arXiv:2607. 16348v1 Announce Type: cross Abstract: Machine learning-based intrusion detection systems (IDSs) often suffer from class imbalance and vulnerability to adversarial attacks, leading to degraded detection performance and reduced robustness.
arXiv:2607. 16214v1 Announce Type: cross Abstract: Image descriptions represented with language models (LMs) predict human brain responses to naturalistic images in high-level visual regions, but the factors driving this predictivity remain unclear.
arXiv:2607. 16240v1 Announce Type: cross Abstract: Direct Alignment Algorithms (DAAs) such as DPO have become a common way to post-train and align LLMs with human preferences.
arXiv:2607. 16242v1 Announce Type: cross Abstract: Fine-Tuning-as-a-Service (FTaaS) platforms let users train large language models (LLMs) on customized tasks, but this pipeline could erode models' safety alignment.
arXiv:2607. 17508v1 Announce Type: cross Abstract: We introduce Retrieval-Augmented Interpretable Learning (RAIL), a probabilistic meta-learning framework for zero-shot generation of task-specific interpretable models that synthesizes coefficient-space structure from natural-language task descriptions and a memory of previously learned task-specific predictors.
arXiv:2607. 16195v1 Announce Type: new Abstract: We identify a structured confound in Reinforcement Learning from Human Feedback (RLHF).
arXiv:2607. 16652v1 Announce Type: cross Abstract: This position paper argues that claims about explanation stability are scientifically invalid without cross method validation.
arXiv:2607. 16303v1 Announce Type: cross Abstract: Medical Vision-Language Models (Med-VLMs) require reliable reasoning from fine-grained visual evidence, yet existing models can produce plausible clinical answers by relying on language priors or medical templates rather than truly attending to diagnosis-critical regions.
arXiv:2607. 16705v1 Announce Type: cross Abstract: Semi-supervised learning (SSL) is an effective solution for medical image segmentation with limited annotations.
arXiv:2607. 16903v1 Announce Type: cross Abstract: Value-aware AI systems require explicit computational representations of human values (groundings) and their aggregation into value systems in order to align their decisions with ours.
arXiv:2607. 16560v1 Announce Type: new Abstract: We propose a language representation for multimodal data in which any observation, whether image, video, or text, is expressed as a bag of atomic propositions, simple statements about the entities, actions, and relations in a scene.
arXiv:2607. 17279v1 Announce Type: cross Abstract: Recently, text-to-video (T2V) models have been widely deployed, sparking growing concerns over their robustness against jailbreak attacks.
arXiv:2607. 17266v1 Announce Type: cross Abstract: Large language models (LLMs) have demonstrated remarkable capabilities in natural language processing.
arXiv:2607. 16660v1 Announce Type: cross Abstract: The increasing adoption of Large Language Models (LLMs) as AI components in modern software systems introduces distinct security risks to the software supply chain.