About
Research & Development
Unveiling LLM Refusal: A Multi-Directional Approach Using Self-Organizing Maps
TLDR: The paper introduces a novel Multi-Directional (MD) approach to suppress refusal behavior in Large Language Models (LLMs) by leveraging Self-Organizing Maps (SOMs). Unlike...
Research & Development
Bridging Safety Gaps in Large Language Models with Policy Patches
TLDR: A new method called "policy patching" allows for rapid, lightweight safety updates in large language models (LLMs) by adding a small, learnable prefix....
Research & Development
AI Models Begin to Grasp What Makes Math Problems Interesting to Humans
TLDR: A new research paper explores how well Large Language Models (LLMs) and Large Reasoning Models (LRMs) align with human judgments of mathematical problem...
Research & Development
DeepProofLog: A Scalable Approach to Neurosymbolic AI with Efficient Proof Generation
TLDR: DeepProofLog (DPrL) is a novel Neurosymbolic AI system designed to enhance the scalability of logic programming by integrating neural networks to guide proof...
Research & Development
Navigating Energy Markets: The Rise of Probabilistic Price Forecasting
TLDR: This research paper reviews the evolution of probabilistic electricity price forecasting, highlighting the shift from single-value predictions to uncertainty-quantifying methods. It covers techniques...
Research & Development
Understanding, Not Reconstructing: A New Era for Autonomous Driving Planning
TLDR: The research introduces the Tokenized Intent World Model (TIWM), a novel approach to autonomous driving that challenges the need for exhaustive scene modeling....