PaTH Attention: The Next Leap in Context-Aware Language Models Large language models have transformed artificial intelligence, powering everything from chatbots to automated code generation. Yet, even the most advanced models often struggle to follow evolving sta... AI research context awareness large language models machine learning PaTH Attention position encoding reasoning transformers
How MIT’s Adaptive Reasoning Makes AI Smarter and Greener MIT researchers have unveiled a breakthrough method that enables large language models (LLMs) to dynamically manage their computational resources according to the challenge at hand. This leap in effic... adaptive reasoning artificial intelligence energy efficiency large language models machine learning process reward model uncertainty calibration
Whisper Leak: How Encrypted AI Chats Can Still Reveal Your Secrets Many people trust that encrypted messaging with AI chatbots is secure, but recent research from Microsoft challenges this assumption. A newly discovered threat, dubbed Whisper Leak , reveals that even... AI security cybersecurity data privacy encrypted traffic large language models machine learning side-channel attack Whisper Leak
AI Powered Gene Set Analysis: Inside GeneAgent’s Self-Verification Breakthrough GeneAgent is an innovative artificial intelligence agent developed by researchers at the National Institutes of Health (NIH). It is engineered to boost the accuracy of gene set analysis, a critical ta... AI hallucinations AI in genomics biomedical research expert-curated databases gene set analysis large language models self-verifying AI
AlphaEvolve: AI-Powered Mathematical Discovery at Scale Google DeepMind's AlphaEvolve is moving the bar yet again in how artificial intelligence tackles mathematical discovery. Published (preprint) in November 2025 by researchers Bogdan Georgiev, e t al at... AI AlphaProof automated reasoning combinatorics DeepMind evolutionary computation FunSearch large language models machine learning mathematical discovery mathematics optimization Terence Tao
MIT Researchers Are Making AI Text Classifiers More Reliable AI text classifiers are now behind many tools we use daily, from chatbots to content moderation systems. Their accuracy and reliability have become critical but how can you be sure they aren’t easily ... adversarial testing AI large language models machine learning MIT robustness software tools text classification
IBM’s AISteer360 Empowers Safe and Customizable LLM Outputs As large language models (LLMs) become increasingly sophisticated, so does the challenge of ensuring their outputs remain safe, relevant, and trustworthy especially in high-stakes enterprise environme... AI safety AISteer360 AI steering enterprise AI generative AI large language models model alignment open source
Toucan Dataset: Transforming AI Agents Into Digital Doers Toucan, a groundbreaking open-source dataset from IBM and the University of Washington is crafted to propel tool-calling capabilities in large language models (LLMs) to new heights. For AI to move bey... AI agents API integration benchmarking large language models machine learning open source tool-calling Toucan dataset
When AI Agents Misremember: How Fake Memories Put Smart Assistants at Risk What if you entrust your AI assistant with your credit card to book a flight, only to wake up and discover it has spent your money on bizarre purchases? What would you do? Panic? This unsettling possi... AI assistants AI security autonomous agents large language models memory manipulation prompt injection Web3
AI-Powered Plant Materials Discovery: Large Language Models Extract Nature's Engineering Secrets From Ancient Wisdom to Modern Innovation Plants have spent hundreds of millions of years perfecting sophisticated engineering solutions that humans are only beginning to understand. From the elegant m... artificial intelligence bio-inspired design biomimetics large language models materials discovery plant materials retrieval augmented generation sustainability
MIT's CodeSteer Helps Language Models Outsmart Complex Problems Large language models (LLMs) have dramatically changed our relationship with AI, offering impressive fluency in language understanding and generation. Yet, when these models confront tasks that demand... AI coaching algorithmic tasks artificial intelligence code generation large language models machine learning MIT research symbolic reasoning
AMD Ryzen AI Max+ Upgrade: Powering 128B-Parameter LLMs Locally on Windows PCs With AMD's latest update deploying massive language models, up to 128 billion parameters, directly on your Windows laptop is now a possible. AMD’s Ryzen AI Max+ is a breakthrough that brings state-of-... AMD context window large language models LLM deployment local AI quantization Ryzen AI Windows AI