Researchers propose a new method called SafeSteer to align large language models with human values without degrading their general capabilities. They ...
WorldLines is a new benchmark for long-horizon stateful embodied agents. These agents must remember user routines and past interactions to assist huma...
Researchers have developed a new AI system called AgRefactor that can automatically convert real-world software into code compatible with High-Level S...
Researchers have found that post-training quantization, a method used to compress large language models for deployment on resource-constrained devices...
Researchers have found that large language models can be vulnerable to 'jailbreaks' where seemingly safe instructions become unsafe when executed in t...
Researchers from Ye Chen and Weining Zhang have proposed a new approach to evaluating large language models (LLMs). They compared two strategies for s...
Researchers have developed a new world model called MemWM that uses memory to improve predictions of how environment states change in response to agen...
A new AI system called TRACE-Memory aims to improve personalized generation by selectively using user history. It does this in two stages: first, it i...
Researchers propose a new evaluation framework called Q-CARE for Retrieval-Augmented Generation (RAG) models. The framework assesses the performance o...
Researchers propose an Inverse Theory of Mind (IToM) pipeline to better understand user behavior in dynamic interfaces. The IToM pipeline infers users...
Researchers have revisited the concept of preprocessing invariance in spectral foundation models, specifically Raman foundation models. They found tha...
Researchers have proposed a new approach to computing market prices and allocations that takes into account negative externalities such as the cost of...