·
DataBubble
  • Home
  • Models
  • News
  • Compare
  • Boards
  • Pricing
  • About
  • Newsletter
  • Methodology
  • Contact
Latest
Integrating High-Level Requirements to Low-Level Tests with Machine-Readable V&V Specifications4m◆Lifelong Multi-Subsystem Pickup and Delivery with Buffer-Limited Handover Stations4m◆MXSens: Sensitivity-Aware Mixed-Precision Quantization for Efficient LLM Inference4m◆Mobile Network Control with a World Model4m◆DA-Fusion: Deformable Attention-Based RGB-D Fusion Transformer for Unseen Object Instance Segmentation4m◆Seg2Grasp: A Robust Modular Suction Grasping in Bin Picking4m◆Generalize and Guide: Decomposing Rewards for Few-Shot Inverse Reinforcement Learning4m◆Time-Frequency Consistency Learning for Robust Speech Deepfake Detection4m◆Autonomous Discovery of Wireless Communications Algorithms4m◆FIFA World Cup 2026 as a Contamination-Free Benchmark for LLM Forecasting Agents: Four Models, a Bookmaker, and 104 Matches4m◆Measuring Monosemanticity in Sparse Autoencoders via Latent Activation Coherence4m◆Persona-as-Configuration: Generative Stakeholder Reporting for Agricultural Floods4m◆CDIS: Cross-Dimensional Class-Agnostic 3D Instance Segmentation via 2D Mask Tracking and 3D-2D Projection Merging4m◆ETAS: An Effect-Typed Language for Agent Systems4m◆BrainNext: A General-Purpose Self-Supervised Foundation Model for Brain MRI Analysis4m◆Feature Attribution-Based Explainability Analysis of Deep Learning Models in Predictive Process Monitoring4m◆Reasoning as a Double-Edged Sword: Architecture and Cross-Stage Robustness in Vision-Language-Action Models4m◆Medical Imaging Fusing Vision Transformer: Laryngeal Cancer Screening with Explanation4m◆ReViV: Reconstructing the Viewer and the View in 4D from Monocular Egocentric Video4m◆Vis2Reg: Visibility-Aware Landmark-Free Geometric 3D--2D Registration for Liver Laparoscopy4m◆Integrating High-Level Requirements to Low-Level Tests with Machine-Readable V&V Specifications4m◆Lifelong Multi-Subsystem Pickup and Delivery with Buffer-Limited Handover Stations4m◆MXSens: Sensitivity-Aware Mixed-Precision Quantization for Efficient LLM Inference4m◆Mobile Network Control with a World Model4m◆DA-Fusion: Deformable Attention-Based RGB-D Fusion Transformer for Unseen Object Instance Segmentation4m◆Seg2Grasp: A Robust Modular Suction Grasping in Bin Picking4m◆Generalize and Guide: Decomposing Rewards for Few-Shot Inverse Reinforcement Learning4m◆Time-Frequency Consistency Learning for Robust Speech Deepfake Detection4m◆Autonomous Discovery of Wireless Communications Algorithms4m◆FIFA World Cup 2026 as a Contamination-Free Benchmark for LLM Forecasting Agents: Four Models, a Bookmaker, and 104 Matches4m◆Measuring Monosemanticity in Sparse Autoencoders via Latent Activation Coherence4m◆Persona-as-Configuration: Generative Stakeholder Reporting for Agricultural Floods4m◆CDIS: Cross-Dimensional Class-Agnostic 3D Instance Segmentation via 2D Mask Tracking and 3D-2D Projection Merging4m◆ETAS: An Effect-Typed Language for Agent Systems4m◆BrainNext: A General-Purpose Self-Supervised Foundation Model for Brain MRI Analysis4m◆Feature Attribution-Based Explainability Analysis of Deep Learning Models in Predictive Process Monitoring4m◆Reasoning as a Double-Edged Sword: Architecture and Cross-Stage Robustness in Vision-Language-Action Models4m◆Medical Imaging Fusing Vision Transformer: Laryngeal Cancer Screening with Explanation4m◆ReViV: Reconstructing the Viewer and the View in 4D from Monocular Egocentric Video4m◆Vis2Reg: Visibility-Aware Landmark-Free Geometric 3D--2D Registration for Liver Laparoscopy4m◆
News/kNNGuard: Turning LLM Hidden Activations into a Training-Free Configurable Guardrail
arxiv
PublishedJuly 3, 2026 at 4:00 AM
—neutral

kNNGuard: Turning LLM Hidden Activations into a Training-Free Configurable Guardrail

Source
arxiv.orgfull article ↗
Read on arxiv→
Publisher summary· verbatim

arXiv:2607.02072v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly deployed in domains requiring guardrails to detect unsafe, off-topic, or adversarial prompts. Existing guardrails predominately rely on fine-tuning to build classifiers, which often suffer from low genera

Stay posted· Newsletter

A 5-min weekly brief — top movers, price watch, story of the week.

// no spam · unsubscribe one-click · free forever

Discussion
Source
↗
arxiv
Read original ↗All from arxiv →

No replies yet. Be first.

Source
↗
arxiv
Read original ↗All from arxiv →

Related coverage

More from ARXIV
arxivIntegrating High-Level Requirements to Low-Level Tests with Machine-Readable V&V Specifications4marxivLifelong Multi-Subsystem Pickup and Delivery with Buffer-Limited Handover Stations4marxivMXSens: Sensitivity-Aware Mixed-Precision Quantization for Efficient LLM Inference4marxivMobile Network Control with a World Model4m
The Bubble Brief
WEEKLY

Read AI insights every Tuesday — top movers, new releases, story of the week.

// no spam · unsubscribe one-click · free forever

Originally published on arxiv ↗
HomeModelsNews