Refuse without Refusal: A Structural Analysis of Safety-Tuning Responses for Reducing Fal…
Explorar
Noticias de IA
37834 elementos — filtrados, clasificados y sin duplicados
Knowing What Not to Answer: Selective Non-Compliance in Vision-Language Models
Adaptation Interfaces for In-Context Tabular Foundation Models in Time-to-Event Prediction
Training-Free Halving of Activated Experts in Fine-Grained Mixture-of-Experts Models
Dynamic Adaptation of the LLM Context for Generating Routines with Coupled Semantics
When Do Internal Probes Beat Reading the Answer? Miscalibrated Readouts and Behavior-Conc…
Atlas: Optimizing Deployment of Compound AI Workflows on Heterogeneous Clusters
Repeat-After-Me: Black-Box Adaptive Visual Prompt Injection
Tracing Audio Grounding and Answer Selection in Audio LLMs
Phase Transition Frequency as a Training Time Predictor of Test Accuracy in ResNets
Hakken: Predicting future discoveries to fill the gaps in today's knowledge
Beyond Code Generation: Reliability, Verification, and Cost Economics in the Agentic Soft…
A Systematic Evaluation of Cross-Lingual Consistency Enhancement Methods in Multilingual …
Patterns of Priming in Production: Lexical, Semantic and Structural Alignment in Language…
A Roadmap for MEG Foundation Models
What Moves? Localized Motion Representations for Compositional Scene Control
AlcaTRAz - Anchored Tree-Rule Defense Against Jailbreaks
Scalable Context Orchestration for Serving LLMs Over Voice
Evidence Integration in Large Language Models
A Deep Generative Model for Synthesizing Labeled Wireless Signals
Enhancing Multimodal Emotion Recognition via Multi-Feature Encoding and Attention-Based F…
Trace2Tower: Transition-Aware EigenTrace Induction of Multi-Level Skills for LLM Agents
Abstraction Agent
Commonsense Reasoning in Computer Vision: Foundations, Recent Advancements, and Future Di…
Necessary or Sufficient? Evaluating LLM Explanations With Behavioural Evidence
Cultural Misalignment in Large Language Models: Detection, Measurement, and Mitigation Th…
CABAL: Multi-Agent Simulacra for Tracing the Effects of Collusive Bidding in Peer Review
ACE: Adaptive Calibration-Free Expert Skipping for MoE-based LLMs
Substrate-Aware AI Agents: Execution Context as a First-Class Input
Building a research-software catalog with a coding agent: from hackathon prototype to pub…