What is Lebanon's Beaufort Castle? Why is Israel attacking Nabatieh? Two people have been injured in an Israeli drone strike on a pick-up truck in southern Lebanon, according to state-run media, the latest attack despite a United States-brokered framework agreement intended to pave the way for a phased Israeli withdrawal. The drone struck the vehicle as it was unloading garbage on the outskirts of the towns of Choukine and Kfar Dajjal in the Nabatieh district early on Friday, Lebanon's National News Agency (NNA) said. Later, NNA reported drones also targeted the towns of Kfar Reman and Nabatieh al-Fawqa.

A Ghanaian social media influencer known as Abu Trica, whose real name is Frederick Kumi, has been extradited to the US to face trial for allegedly running a romance scam that defrauded elderly Americans of over $8m (£5.9m). He denies all the charges against him. Prosecutors said he used AI tools to create fake online identities, targeting victims through social media and dating sites, earning their trust then extorting their money. Kumi was flown to the US on Thursday and faces up to 20 years in prison if found guilty of conspiracy to commit wire fraud and money laundering. Kumi's lawyer, Oliver Barker-Vormawor, told the BBC he went to court on Thursday to stop the extradition before learning a short while later that Kumi had in fact already been extradited on board a Delta Airlines flight.

Things to Do in L.A. Tap to enable a layout that focuses on the article. Alberto Carvalho addresses a press conference at Elysian Heights Elementary Arts Magnet in 2022. This is read by an automated voice. Please report any issues or inconsistencies here . See more from the L.A. Times in Google Search.

Plus: Former national security advisor John Bolton pleads guilty in classified-materials case, Microsoft helps take down major infostealer infrastructure, and more. A WIRED investigation this week offers insight into a predictive policing program in Bristol, England that has involved 23 separate models over more than a decade, intended to score the likelihood of specific individuals will perpetrate or be victims of different crimes. The investigation draws on data from public records requests and other reporting to reveal a messy law enforcement apparatus that has real implications for the community--but that most people in the area know nothing about. After the identities of members of Peter Thiel's private "Dialog" group were exposed last week, the organization claimed that a "criminal" hacker was behind the breach. But evidence shows that members' personal information--including that of a White House intelligence official and an active-duty special operations officer --was publicly accessible and likely exposed as the result of a Dialog website misconfiguration .

Alignment faking in large language models presented a demonstration of Claude 3 Opus and Claude 3.5 Sonnet selectively complying with a helpfulonly training objective to prevent modification of their behavior outside of training. We expand this analysis to 25 models and find that only 5 (Claude 3 Opus, Claude 3.5 Sonnet, Llama 3 405B, Grok 3, Gemini 2.0 Flash) comply with harmful queries more when they infer they are in training than when they infer they are in deployment. First, we study the motivations of these 5 models. Results from perturbing details of the scenario suggest that only Claude 3 Opus's compliance gap is primarily and consistently motivated by trying to keep its goals. Second, we investigate why many chat models don't fake alignment. Our results suggest this is not entirely due to a lack of capabilities: many base models fake alignment some of the time, and post-training eliminates alignment-faking for some models and amplifies it for others.We investigate 5 hypotheses for how post-training may suppress alignment faking and find that variations in refusal behavior may account for a significant portion of differences in alignment faking.

Large language models are increasingly used to support high-stakes decisions, potentially influencing who is granted bail or receives a loan. Naive chain-ofthought sampling can improve average decision accuracy, but has also been shown to amplify unfair bias. To address this challenge and enable the trustworthy use of reasoning models in high-stakes decision-making, we propose a framework for training a generalizable Fairness Reward Model (FRM). Our model assigns a fairness score to LLM reasoning, enabling the system to down-weight biased trajectories and favor equitable ones when aggregating decisions across reasoning chains. We show that a single Fairness Reward Model, trained on weakly supervised, LLM-annotated examples of biased versus unbiased reasoning, transfers across tasks, domains, and model families without additional fine-tuning. When applied to real-world decision-making tasks including recidivism prediction and social media moderation, our approach consistently improves fairness while matching, or even surpassing, baseline accuracy.

A freelance HR consultant paid Garfield AI about £400 to send a legal letter and then issue court proceedings over an unpaid debt of £7,000. A freelance HR consultant paid Garfield AI about £400 to send a legal letter and then issue court proceedings over an unpaid debt of £7,000. Barrister who was given material produced by Garfield AI says advocacy at trial'remained fundamentally human' An artificial intelligence law firm has won a case in an English court, in what is believed to be the first time a trial has been won using an AI lawyer. A freelance HR consultant, Tamires Camal Taquidir, paid the firm, Garfield AI, about £400 to send a legal letter and then issue court proceedings over an unpaid debt of £7,000. The co-founder of Garfield, Philip Young, called it a "landmark moment" for access to justice and said many small businesses have had to write off debts because the cost of litigation outweighed the money they could hope to win.

The increasing deployment of Large Vision-Language Models (LVLMs) raises safety concerns under potential malicious inputs. However, existing multimodal safety evaluations primarily focus on model vulnerabilities exposed by static image inputs, ignoring the temporal dynamics of video that may induce distinct safety risks. To bridge this gap, we introduce Video-SafetyBench, the first comprehensive benchmark designed to evaluate the safety of LVLMs under video-text attacks.

Reinforcement learning (RL) has become a predominant technique to align language models (LMs) with human preferences or promote outputs which are deemed to be desirable by a given reward function. Standard RL approaches optimize average reward, while methods explicitly focused on reducing the probability of undesired outputs typically come at a cost to average-case performance. To improve this tradeoff, we introduce RePULSe, a new training method that augments the standard RL loss with an additional loss that uses learned proposals to guide sampling low-reward outputs, and then reduces those outputs' probability. We run experiments demonstrating that RePULSe produces a better tradeoff of expected reward versus the probability of undesired outputs and is more adversarially robust, compared to standard RL alignment approaches and alternatives.

Text-to-Image (T2I) models have achieved remarkable success in generating visual content from text inputs.