Records found: 6
REST is a new evaluation framework that tests large reasoning models with multiple problems simultaneously, revealing performance drops and providing deeper insights into their real-world multitasking abilities.
A new study reveals that large reasoning models, while powerful, expose sensitive information through their reasoning traces, highlighting significant privacy risks in AI personal assistants.
AREAL is a new asynchronous reinforcement learning system that significantly speeds up training of large reasoning models by separating generation and training processes, achieving up to 2.77× faster training without loss of accuracy.
Apple researchers have uncovered fundamental weaknesses in large reasoning AI models through controlled puzzle evaluations, showing significant performance drops as task complexity increases.
‘DARPA’s expMath initiative and advanced AI models are accelerating mathematical research, helping solve problems once thought intractable while still relying on human creativity for true breakthroughs.’
WebThinker is a new AI agent that empowers large reasoning models to autonomously search the web and generate detailed scientific reports, significantly improving performance on complex reasoning benchmarks.