Alignment, interpretability, red-teaming, bias and privacy: the research on what these systems do when they misbehave.
arXiv:2610.07177v1 Announce Type: cross Abstract: An open contrastive decision model is near chance as a judge on the hard public benchmarks: Contrastive-LM/CLM-v0.1-8B scores between 0.351 (best- of...
The paper proposes the Analytical Memory Unit (AMU), a memory schema that attaches a full derivation (lineage) graph to every cached result in enterprise AI agents. By gating retrieval with a policy that requires authorization for every column touched, the authors prove that sensitive columns cannot be leaked through derived results, achieving up to 90% lineage completeness to eliminate leakage. Experiments show lineage‑gated retrieval removes 18.8‑25.5% of cross‑department leakage while maintaining 81.5‑82.6% memory reuse with minimal overhead, and a real‑agent proof‑of‑concept demonstrates zero leaks over multiple interactions.
arXiv:2610.07276v1 Announce Type: cross Abstract: Deployment-time safety of language models is commonly implemented through runtime guardrails such as input moderation, routing, retrieval verificatio...
arXiv:2610.07388v1 Announce Type: cross Abstract: Joint Models simultaneously model longitudinal and survival outcomes, leveraging patterns in patients' longitudinal trajectory to improve the predict...
arXiv:2610.07476v1 Announce Type: cross Abstract: Frontier AI treaties or agreements on limiting computation require external verification; an external auditor must be able to confirm how much comput...
arXiv:2610.07532v1 Announce Type: cross Abstract: LLMs have advanced rapidly, raising growing concerns about their safety. Recent work has proposed approaches to detect and defend against attacks inc...
The paper introduces Physics‑Guided Visual Prompting (PG‑VP), a plug‑and‑play module that overlays a virtual obstacle onto the input of a frozen Vision‑Language‑Action model to guide navigation around invisible hazards such as radiation or temperature spikes. PG‑VP performs a physics‑based risk assessment to determine the avoidance direction and dynamically renders the same virtual obstacle across frames, allowing the existing navigation policy to detour without retraining. Experiments on OmniNav with R2R‑CE and RxR‑CE datasets show that PG‑VP steers the policy toward low‑risk actions in 84.9% and 83.2% of cases, while real‑world tests on a robot demonstrate significant safety improvements against thermal and radiation sources.
arXiv:2610.07583v1 Announce Type: cross Abstract: While AI weather models now rival operational forecasts, how they represent the atmosphere internally remains an open question: feature attribution r...
arXiv:2610.07654v1 Announce Type: cross Abstract: On-policy distillation (OPD) has attracted growing attention as an effective way to transfer capabilities from teacher models to student models. Rece...
arXiv:2610.08093v1 Announce Type: cross Abstract: Developing reliable models for clinical tasks, such as Medical Question Answering (QA), is severely constrained by the limited availability of high-q...
The paper investigates how acoustic and content-oriented speaker verification attacks perform against multilingual voice anonymization. It evaluates both attacker types on anonymized speech, builds a multilingual voice-converted dataset to enhance cross‑lingual generalization, and finds that attack success depends on how much linguistic utility is preserved. Acoustic‑oriented attacks generally outperform content‑oriented ones, but the performance gap narrows when linguistic information is well maintained, and the new dataset improves results while reducing cross‑lingual disparities.
arXiv:2610.08316v1 Announce Type: cross Abstract: Protein Generative Models (PGMs) have revolutionized structural biology by enabling the design of complex 3D protein structures from sequence data. H...
arXiv:2610.08331v1 Announce Type: cross Abstract: The rapid integration of Vision Language Models (VLMs) into sensitive systems introduces critical safety vulnerabilities that remain unexplored in ex...
arXiv:2610.08388v1 Announce Type: cross Abstract: Large language models (LLMs) have demonstrated strong capabilities in question answering, yet they still frequently suffer from hallucinations on kno...
arXiv:2610.08448v1 Announce Type: cross Abstract: On-Policy Distillation (OPD) trains a student on its own generations using teacher feedback. With different tokenizers, comparing teacher and student...
The paper argues that probe scores lack intrinsic meaning and should be interpreted relative to two reference points: a floor (what simple inputs predict) and a ceiling (what the full input predicts). The difference, called headroom, indicates the range where a probe can reveal that a model computes beyond what the input already provides. Experiments on transformers and real models show that headroom can vanish when the target no longer depends on hidden variables or when the input no longer reveals them, and that some previously claimed representations are largely explained by the input text alone.
The paper systematically reviews 100 empirical HCI studies on children and youth interacting with AI in various settings. Using the YAIR risk taxonomy and MIT Mitigation Taxonomy, it maps identified risks to countermeasures, finding that most risks are only matched with proposed countermeasures, few are implemented, and even fewer are evaluated. The authors highlight gaps in coverage and testing, and suggest concrete research directions to improve youth AI safety.
arXiv:2610.08559v1 Announce Type: cross Abstract: Activation steering has gained popularity as a lightweight inference-time debiasing technique for large language models. However, prior work reports...
FedDermaSeg explores federated learning for skin lesion segmentation, aiming to preserve privacy by avoiding centralized data aggregation. Using the ISIC 2018 dataset in a simulated distributed setting, the federated model matches centralized training performance and outperforms locally trained models. The study demonstrates that collaborative, privacy‑preserving segmentation is feasible without central image storage.
arXiv:2610.08760v1 Announce Type: cross Abstract: Recent advances in world models have enabled increasingly realistic visual synthesis. However, these generated environments remain largely silent. Br...