Anthropic Investigates How Large Language Models Develop a Character
Recent research by Anthropic engineers explores identifiable patterns of activity that seems to give rise to an emerging personality. These traits, known as persona vectors, help explain how a model’s personality shifts over its lifecycle and lay the groundwork for better controlling those changes.
Roblox Open-sources AI System to Detect Conversations Potentially Harmful to Kids
Roblox Sentinel is an AI system designed to detect early signs of potential child endangerment for further analysis and investigation. Implemented as a Python library, Sentinel uses contrastive learning to handle highly imbalanced datasets that often challenge traditional classifiers and can be applied to a wide range of use cases.
Gemini 2.5 Deep Think Parallelizes Creative Problem-Solving
As part of Google AI Ultra subscription, Gemini 2.5 Deep Think is a model designed for creative problem-solving through the use of parallel thinking techniques and extended inference time.
IBM and NASA Trained the First Foundational Model for Heliophysics
Surya is a 366M-parameter model created by IBM and NASA to forecast solar activity, including flare events, solar winds, and precursors to solar eruptions, which can significantly impact astronaut safety in space as well as terrestrial systems such as communications, power distribution, and more.