About
Research & Development
Strengthening AI Safety: Distributing Protective Mechanisms Across Language Models
TLDR: Current AI safety measures in large language models (LLMs) are vulnerable because they rely on a limited number of internal components called 'attention...
Research & Development
Unlocking Commonsense in Compact Vision-Language Models
TLDR: A new framework, NLKI, significantly improves small Vision-Language Models (sVLMs) in commonsense visual question answering by integrating natural language facts and LLM-generated explanations....
Research & Development
Advancing Multi-Robot Navigation with Discrete-Guided Diffusion
TLDR: The paper introduces Discrete-Guided Diffusion (DGD), a novel framework for Multi-Robot Motion Planning (MRMP). DGD combines discrete Multi-Agent Path Finding (MAPF) solutions with...
Research & Development
Unmasking Truth: How Quantization Affects LLM Honesty
TLDR: A new study introduces TruthfulnessEval, a framework to assess the honesty of quantized Large Language Models (LLMs). It finds that while quantized LLMs...
Research & Development
Unmasking Covert Actions: How to Reliably Monitor Autonomous AI
TLDR: This paper introduces Monitor Red Teaming (MRT), a framework to test systems designed to detect hidden misbehavior in LLM agents. It finds that...
Research & Development
Automating Course Material Creation with Collaborative AI Agents
TLDR: Instructional Agents is a new multi-agent large language model (LLM) framework designed to automate the end-to-end generation of high-quality course materials, including syllabi,...