AI
The Illusion of Equivalency: Statistical Characterization of Quantization Effects in LLMs
Researchers have found that post-training quantization, a method used to compress large language mod...
AI
When Words Are Safe But Actions Kill: Probing Physical Jailbreak Beyond Textual Jailbreak in Hidden-State Risk Space
Researchers have found that large language models can be vulnerable to 'jailbreaks' where seemingly ...
AI
Share the Judge, Learn the Deferral: Where Specialization Helps LLM Evaluation
Researchers from Ye Chen and Weining Zhang have proposed a new approach to evaluating large language...
AI
MemWM: Memory-Augmented Text-Based World Model
Researchers have developed a new world model called MemWM that uses memory to improve predictions of...
AI
TRACE-Memory: Public-Conditioned Retrieval and Utility-Aware Evidence Admission for Personalized Generation
A new AI system called TRACE-Memory aims to improve personalized generation by s...
AI
Towards Query-Agnostic RAG Evaluation via Query Coverage and Claim Verifiability
Researchers propose a new evaluation framework called Q-CARE for Retrieval-Augme...
AI
Inverse Theory of Mind Modeling for Content Recommendation: From Web Browsing to Dynamic Intelligent Interfaces
Researchers propose an Inverse Theory of Mind (IToM) pipeline to better understa...
AI
Attributing Preprocessing Invariance in Spectral Foundation Models
Researchers have revisited the concept of preprocessing invariance in spectral f...