This paper introduces STAR-1, a high-quality, just-1k-scale safety dataset specifically designed for large reasoning models (LRMs) like DeepSeek-R1. Built on three core principles -- diversity, deliberative reasoning, and rigorous filtering -- STAR-1 aims to address the critical needs for safety alignment in LRMs. Specifically, we begin by integrating existing open-source safety datasets from diverse sources. Then, we curate safety policies to generate policy-grounded deliberative reasoning samples. Lastly, we apply a GPT-4o-based safety scoring system to select training examples aligned with best practices. Experimental results show that fine-tuning LRMs with STAR-1 leads to an average 40% improvement in safety performance across four benchmarks, while only incurring a marginal decrease (e.g., an average of 1.1%) in reasoning ability measured across five reasoning tasks. Extensive ablation studies further validate the importance of our design principles in constructing STAR-1 and analyze its efficacy across both LRMs and traditional LLMs. Our project page is https://ucsc-vlaa.github.io/STAR-1.

Large language models (LLMs) increasingly operate in social contexts, motivating analysis of how they express and shift moral judgments. In this work, we investigate the moral response of LLMs to persona role-play, prompting a LLM to assume a specific character. Using the Moral Foundations Questionnaire (MFQ), we introduce a benchmark that quantifies two properties: moral susceptibility and moral robustness, defined from the variability of MFQ scores across and within personas, respectively. We find that, for moral robustness, model family accounts for most of the variance, while model size shows no systematic effect. The Claude family is, by a significant margin, the most robust, followed by Gemini and GPT-4 models, with other families exhibiting lower robustness. In contrast, moral susceptibility exhibits a mild family effect but a clear within-family size effect, with larger variants being more susceptible. Moreover, robustness and susceptibility are positively correlated, an association that is more pronounced at the family level. Additionally, we present moral foundation profiles for models without persona role-play and for personas averaged across models. Together, these analyses provide a systematic view of how persona conditioning shapes moral behavior in large language models.

Large Language Models (LLMs) achieve excellent performance in natural language reasoning tasks through pre-training on vast unstructured text, enabling them to understand the logic in natural language and generate logic-consistent responses. However, the representational differences between unstructured and structured knowledge make LLMs inherently struggle to maintain logic consistency, leading to \textit{Logic Drift} challenges in structured knowledge reasoning tasks such as Knowledge Graph Question Answering (KGQA). Existing methods address this limitation by designing complex workflows embedded in prompts to guide LLM reasoning. Nevertheless, these approaches only provide input-level guidance and fail to fundamentally address the \textit{Logic Drift} in LLM outputs. Additionally, their inflexible reasoning workflows cannot adapt to different tasks and knowledge graphs. To enhance LLMs' logic consistency in structured knowledge reasoning, we specifically target the logits output from the autoregressive generation process. We propose the \textit{Logits-to-Logic} framework, which incorporates logits strengthening and logits filtering as core modules to correct logical defects in LLM outputs. Extensive experiments show that our approach significantly improves LLMs' logic consistency in structured knowledge reasoning and achieves state-of-the-art performance on multiple KGQA benchmarks.

Oscar-winning actors Matthew McConaughey and Michael Caine have both signed a deal with the AI audio company ElevenLabs. The New York-based company can now create AI-generated versions of their voices as part of a bid to solve a "key ethical challenge" in the artificial intelligence industry's alliance with Hollywood. Emma Thompson speaks of her'intense irritation' with AI McConaughey, who has also invested in the company and collaborated with it since 2022, will now allow ElevenLabs to translate his newsletter, Lyrics of Livin', into a Spanish-language audio version using his voice. In a statement, the Dallas Buyers Club actor said he was "impressed" by ElevenLabs and wanted the partnership to help him "reach and connect with even more people". ElevenLabs is also launching the Iconic Voices Marketplace, which will allow brands to partner with the company and use officially licensed celebrity voices for AI-generated usage.

Fraudsters are targeting veterans with fake VA overpayment scams through texts, emails and calls demanding urgent payments via gift cards or wire transfers.

The Turtle Conservancy is rescuing reptiles, while preserving their precious DNA. Breakthroughs, discoveries, and DIY tips sent every weekday. Nestled in rural New Jersey's rolling hills, a top secret animal sanctuary is keeping its occupants safe. The oasis is not for the official state animal (the horse) or even birds rescued from the Jersey shore. This hidden refuge belongs to turtles and tortoises.

It helps pilots sense how the vehicle is performing and adds a cinematic touch to the flying experience. It's part science, part art and a hint at how human-centered this design really is. Future pilots can train in a realistic Sigma cockpit simulator and even log hours toward a license.

This material may not be published, broadcast, rewritten, or redistributed. Quotes displayed in real-time or delayed by at least 15 minutes. Market data provided by Factset . Powered and implemented by FactSet Digital Solutions . Mutual Fund and ETF data provided by Refinitiv Lipper .

Trump boasts Republicans'broke' Chuck Schumer during shutdown slugfest amid growing calls from the far-left to remove minority leader Camp Mystic accused of neglect in'self-created disaster' after 25 young campers die in flood Gavin Newsom just let slip the Democrats' dirty secret that could totally destroy them... and Trump is loving it: MARK HALPERIN You think you look so youthful and trendy. But these are the brutal facts about your shoes, hats, jumpers and denim. You're being lied to - this is the truth I'm a heart surgeon... here is what you must NEVER do after turning 40 Paris Jackson reveals she has a HOLE in her nose from drug abuse... after celebrating five years of sobriety Warren Buffett's final farewell: the billionaire who once fingerprinted nuns says he's'going quiet' Shamed Allison Mack's husband revealed as former neo-Nazi Kim Kardashian flaunts her famous curves in daring cut-out gown at All's Fair premiere in Brazil as she's seen for first time since failing the bar exam Timothee Chalamet has DUMPED Kylie Jenner, insiders claim... and reveal brutal new way he is humiliating her Trump dismisses economic anxiety with'fake' polls and reveals reason 600,000 Chinese students are in US Was ANOTHER person inside the house with'Foxy Knoxy' when Meredith Kercher was killed 18 years ago? That's the sensational new claim of the Italian prosecutor who put the American student in jail, DAVID JONES reveals Disgraceful lies of'everyday' mom exposed after she chased down ICE agents in her Ford Mustang then cried wolf Millennial socialist who still lives off parents' money is set to become next Seattle mayor Netflix's historical drama Death By Lightning branded woke after fans spot bizarre detail... can you see it? Revealed: MAGA Trump-supporting influencer and friend are subject of police photo appeal over alleged'assault and racial abuse' in London AMANDA PLATELL: It's time the royals stripped Beatrice and Eugenie of their titles.

Trump boasts Republicans'broke' Chuck Schumer during shutdown slugfest amid growing calls from the far-left to remove minority leader Camp Mystic accused of neglect in'self-created disaster' after 25 young campers die in flood Gavin Newsom just let slip the Democrats' dirty secret that could totally destroy them... and Trump is loving it: MARK HALPERIN You think you look so youthful and trendy. But these are the brutal facts about your shoes, hats, jumpers and denim. You're being lied to - this is the truth I'm a heart surgeon... here is what you must NEVER do after turning 40 Paris Jackson reveals she has a HOLE in her nose from drug abuse... after celebrating five years of sobriety Warren Buffett's final farewell: the billionaire who once fingerprinted nuns says he's'going quiet' Shamed Allison Mack's husband revealed as former neo-Nazi Kim Kardashian flaunts her famous curves in daring cut-out gown at All's Fair premiere in Brazil as she's seen for first time since failing the bar exam Timothee Chalamet has DUMPED Kylie Jenner, insiders claim... and reveal brutal new way he is humiliating her Trump dismisses economic anxiety with'fake' polls and reveals reason 600,000 Chinese students are in US Was ANOTHER person inside the house with'Foxy Knoxy' when Meredith Kercher was killed 18 years ago? That's the sensational new claim of the Italian prosecutor who put the American student in jail, DAVID JONES reveals Disgraceful lies of'everyday' mom exposed after she chased down ICE agents in her Ford Mustang then cried wolf'Grand conspiracy' probe seeks to prove cabal of'deep state' Democrats led by Obama tried to destroy Trump Taylor Swift's bridesmaids revealed as A-listers are tapped to join her wedding squad ahead of Travis Kelce nuptials Millennial socialist who still lives off parents' money is set to become next Seattle mayor Netflix's historical drama Death By Lightning branded woke after fans spot bizarre detail... can you see it? New York City's tallest men reveal the downsides to being over 7ft... from brutal injuries to dating struggles Gen Alpha is cancelling the KEYBOARD: Youngsters won't ever have to write emails when they join the workforce - and will send voice notes to their boss instead, report reveals READ MORE: Microsoft's paperclip mascot Clippy delights users as it returns From floppy disks to fax machines, many once-common office technologies would be baffling to the younger generation.