Safety-aligned large language models (LLMs) are becoming increasingly widespread, especially in sensitive applications where fairness is essential and biased outputs can cause significant harm. However, evaluating the fairness of models is a complex challenge, and approaches that do so typically utilize standard question-answer (QA) styled schemes. Such methods often overlook deeper issues by interpreting the model's refusal responses as positive fairness measurements, which creates a false sense of fairness. In this work, we introduce the concept of silenced biases, which are unfair preferences encoded within models' latent space and are effectively concealed by safety-alignment. Previous approaches that considered similar indirect biases often relied on prompt manipulation or handcrafted implicit queries, which present limited scalability and risk contaminating the evaluation process with additional biases. We propose the Silenced Bias Benchmark (SBB), which aims to uncover these biases by employing activation steering to reduce model refusals during QA. SBB supports easy expansion to new demographic groups and subjects, presenting a fairness evaluation framework that encourages the future development of fair models and tools beyond the masking effects of alignment training. We demonstrate our approach over multiple LLMs, where our findings expose an alarming distinction between models' direct responses and their underlying fairness issues.

We present a suite of packages in R, Python, Julia, and C++ that efficiently solve the Sorted L-One Penalized Estimation (SLOPE) problem. The packages feature a highly efficient hybrid coordinate descent algorithm that fits generalized linear models (GLMs) and supports a variety of loss functions, including Gaussian, binomial, Poisson, and multinomial logistic regression. Our implementation is designed to be fast, memory-efficient, and flexible. The packages support a variety of data structures (dense, sparse, and out-of-memory matrices) and are designed to efficiently fit the full SLOPE path as well as handle cross-validation of SLOPE models, including the relaxed SLOPE. We present examples of how to use the packages and benchmarks that demonstrate the performance of the packages on both real and simulated data and show that our packages outperform existing implementations of SLOPE in terms of speed.

Elon Musk, already the world's richest man, is now on the path to becoming its first trillionaire. Tesla's shareholders recently approved a massive pay package for the CEO, including some $1 trillion in stock options. But the payout will happen only if certain targets are met--including Musk's successful deployment of 1 million Optimus robots. Named after a character, because of course it is, Optimus is a humanoid machine that's supposed to be able to complete boring and dangerous work in place of humans. The robot was unveiled in 2021, when Tesla held an "AI Day" event detailing its future plans.

On this episode of, guest Brian Merchant walks us through a historical framework he used to analyze whether AI fits the classic signs of an economic bubble--and what that means for all of us. Chatter about an AI bubble has been everywhere lately, and top tech companies like Google, Meta, and Microsoft have doubled down on their AI investments for 2026. But how have analysts in the past accurately identified forming tech bubbles? Hosts Michael Calore and Lauren Goode sit down with Brian Merchant, WIRED contributor and author of the newsletter to break down the four criteria some researchers have used in the past to understand and brace for the worst. Please help us improve by filling out our listener survey . Write to us at uncannyvalley@wired.com . You can always listen to this week's podcast through the audio player on this page, but if you want to subscribe for free to get every episode, here's how: If you're on an iPhone or iPad, open the app called Podcasts, or just tap this link . Hey Lauren, how are you doing? It's earnings season, so a lot of us on the business desk here at WIRED have been tuning into tech companies earnings reports and their earnings calls. And I guess that basically means it's CapEx season. Now that I'm a business desk reporter, I say CapEx. I throw it around at parties. But we are seeing a trend in how tech companies are sleeping on piles of money, but they aren't just sleeping on it. They're sharing big plans to spend on it, and especially to spend on AI infrastructure. And this is all partly what is fueling all of this talk about a bubble, which we touched on a little bit a couple of weeks ago with our colleague Molly Taft.

Valar Atomics Says It's the First Nuclear Startup to Achieve Criticality A Trump administration pilot program aims for three nuclear startups to reach a key milestone by July 4, 2026. Valar Atomics says it's the first to do so--but it had some help. The El Segundo, California-based startup, which last week announced it had secured a $130 million funding round with backing from Palmer Luckey and Palantir CTO Shyam Sankar, claims that it is the first nuclear startup to create a critical fission reaction. It's also, more specifically, the first company in a special Department of Energy pilot program aiming to get at least three startups to criticality by July 4 of next year to announce it had achieved this reaction. The pilot program, which was formed following an executive order president Donald Trump signed in May, has upended US regulation of nuclear startups, allowing companies to reach new milestones like criticality at a rapid pace.

The long-nosed Pinocchio chameleon is multiple species. Biologists have finally solved the century-old reptilian mystery. Breakthroughs, discoveries, and DIY tips sent every weekday. For nearly 150 years, zoologists have taken the Pinocchio chameleon () at face value.. However, a recent reexamination detailed in reveals that the chameleon is actually multiple species with elongated snouts worthy of the nickname.

Miss Israel Melanie Shiraz said she received death and rape threats after controversial video appears to show her glaring at Palestinian contestant during Miss Universe pageant in Thailand.

In 1996 Nasa and the White House made the explosive announcement that the rock contained traces of Martian bugs. The meteorite, catalogued as Allen Hills (ALH) 84001, crashed onto the frozen wastes of Antarctica 13,000 years ago and was recovered in 1984. Photographs were released showing elongated segmented objects that appeared strikingly lifelike.

Chimpanzees' brutal battle for territory leads to a baby boom A rival chimp can die in less than 15 minutes during these deadly territorial fights. New research led by UCLA and the University of Michigan has shown that chimp communities that kill their neighbors to gain territory also gain reproductive advantages. Breakthroughs, discoveries, and DIY tips sent every weekday. Uganda's Ngogo chimpanzees are well known for their "chimpanzee warfare." Primatologists have observed their brutal, lethal fights between 10 or more chimpanzees for decades, deciphering what leads to such violence.