Safety benchmarks typically report one overall score for a suite of datasets, each of which may target one or more safet…
LLM developers post-train their models to exhibit prosocial values and behavioral traits, which are enumerated in an ali…
Generating realistic human-object interactions (HOI) in complex 3D scenes requires two complementary capabilities: reaso…
A generalist robot should not only perform diverse tasks but also improve through experience, turning what it learns dur…
AI agents can now conduct real-world cyberattacks, scale up capabilities with the number of agents, and collectively pur…
Recent incidents have highlighted the challenge of monitoring LLM agents and the danger of models deceiving people. We s…
Bifurcations are ubiquitous in physical systems, from structural buckling to fluid and climate dynamics, yet they remain…
We propose BrickBench, a benchmark for agentic text-conditioned LEGO-set design. Given a prompt, an agent is tasked with…
In 2026, cybersecurity evaluations involving OpenAI, Anthropic, and Google agents reached real systems outside their aut…
METR's 50\% time horizon measures the human completion time of software tasks that an AI solves with 50\% probability, a…
Blocking a receptor called P2X7 significantly reduced inflammation in human brain tissue, revealing a potential new trea…
Physicists have found striking evidence for an unusual form of superconducting behavior in uranium ditelluride, where el…
Earth’s center of mass is constantly on the move, shifting by a few millimeters as enormous amounts of water, ice, and e…
The James Webb Space Telescope has found strong evidence that the rocky lava world HD 3167 b has an atmosphere, despite…
Scientists at UC San Francisco have developed a way to create cancer-fighting CAR-T cells directly inside the body, pote…
A Stanford-designed molecule turns a major lymphoma-promoting protein against cancer cells, activating their own self-de…
Giving zoo lions whole carcasses on a feast-and-fast schedule appears to bring out more natural behavior. The lions spen…
A study of more than 130,000 babies born in Georgia found that hundreds of fathers died within the first five years of t…
A large analysis of 22 studies found that the trans fats naturally present in dairy foods such as milk, butter, yogurt,…
Just seven strength-training sessions helped obese mice shrink visceral fat cells and reactivate molecular pathways invo…