Alignment, interpretability, red-teaming, bias and privacy: the research on what these systems do when they misbehave.
arXiv:2610.15343v1 Announce Type: cross Abstract: World models and video understanding models are increasingly deployed in dynamic visual environments, yet their robustness to physical adversarial th...
arXiv:2610.15280v1 Announce Type: cross Abstract: Depth estimation and Gaussian reconstruction share a need for multi-scale feature processing, yet use different prediction interfaces and training ob...
arXiv:2610.15513v1 Announce Type: cross Abstract: Safety-aligned large language models (LLMs) remain vulnerable to adversarial prompts that bypass their safety training. Existing automated red-teamin...
arXiv:2610.19763v1 Announce Type: new Abstract: Physics-guided networks for underwater image restoration predict transmission and backscatter and invert the image-formation model, yet rarely test whe...
arXiv:2610.17376v1 Announce Type: new Abstract: A shared latent space is often assumed to make representations from different schemas directly interchangeable, yet sharing parameters, dimensionality,...
arXiv:2610.17473v1 Announce Type: new Abstract: Generative safety guards must classify requests correctly and return outputs that applications can parse. Reasoning guards can improve classification a...
arXiv:2610.17816v1 Announce Type: new Abstract: Edge-cloud collaboration combines the lower inference overhead and greater data locality of edge-side execution with the stronger capabilities of cloud...
arXiv:2407.12703v5 Announce Type: cross Abstract: Fine-tuning pre-trained language models (PLMs) has recently shown a potential to improve knowledge graph completion (KGC). However, most PLM-based me...
arXiv:2610.19968v1 Announce Type: new Abstract: Federated fine-tuning enables large language models (LLMs) to learn collaboratively from distributed client data, but malicious client updates can comp...
arXiv:2610.14071v1 Announce Type: cross Abstract: As LLM agents are increasingly deployed in high-stakes enterprise settings, understanding how they handle sensitive information in agentic dialogue h...
arXiv:2610.19911v1 Announce Type: new Abstract: Adversarial robustness distillation typically treats a robust teacher as a fixed knowledge source and transfers robustness to a compact student in a on...
arXiv:2610.14476v1 Announce Type: cross Abstract: Comparing large language models can reveal shared representational structure and systematic differences between them. However, aggregate similarity m...
arXiv:2610.14473v1 Announce Type: cross Abstract: We propose a new framework for banner ad auctions, leveraging diffusion models to allow for multiple bidders to co-win a single banner ad opportunity...
arXiv:2610.15008v1 Announce Type: new Abstract: Language models are frequently aligned to human preferences via imperfect evaluators whose errors can be exploited during optimization. We train polici...
arXiv:2610.14523v1 Announce Type: new Abstract: ScienceBuddy-PV adapts the ScienceBuddy recursive-in-recursive design to panvascular intervention research through a structured evidence ledger, device...
arXiv:2610.14625v1 Announce Type: new Abstract: Recursive self-improvement (RSI) offers a route to AI systems that participate in their own continued development. Recent methods allow language agents...
arXiv:2610.16312v1 Announce Type: new Abstract: Explicit reasoning in large language models (LLMs) is often associated with reliability, yet it can still pose risks when a model infers harmful goals...
arXiv:2610.17410v1 Announce Type: new Abstract: Advances in large-scale machine learning have enabled agentic AI systems: persistent, autonomous computational entities that pursue goals, initiate act...
arXiv:2610.16916v1 Announce Type: new Abstract: The European Union's Artificial Intelligence Act introduces a structured framework for governing high-risk AI systems, emphasizing transparency, risk m...
arXiv:2610.18158v1 Announce Type: new Abstract: Learning dealer agents can settle on persistently wide quotes, raising what customers pay to trade. A venue can route more orders to tighter quotes, bu...