Aligning LLM-based judges with human preferences is a significant challenge, as they are difficult to calibrate and often suffer from rubric sensitivity, bias, and instability. Overcoming this challenge advances key applications, such as creating reliable reward models for Reinforcement Learning from Human Feedback (RLHF) and building effective routing systems that select the best-suited model for a given user query. In this work, we propose a framework for modeling diverse, persona-based preferences by learning to aggregate outputs from multiple rubric-conditioned judges. We investigate the performance of this approach against naive baselines and assess its robustness through case studies on both human and LLM-judges biases. Our primary contributions include a persona-based method for synthesizing preference labels at scale and two distinct implementations of our aggregator: Generalized Additive Model (GAM) and a Multi-Layer Perceptron (MLP).

ABSTRACT Advancements in spoken language processing have driven the development of spoken language models (SLMs), designed to achieve universal audio understanding by jointly learning text and audio representations for a wide range of tasks. Although promising results have been achieved, there is growing discussion regarding these models' generalization capabilities and the extent to which they truly integrate audio and text modalities in their internal representations. In this work, we evaluate four SLMs on the task of speech emotion recognition using a dataset of emotionally incongruent speech samples, a condition under which the semantic content of the spoken utterance conveys one emotion while speech expressiveness conveys another. Our results indicate that SLMs rely predominantly on textual semantics rather than speech emotion to perform the task, indicating that text-related representations largely dominate over acoustic representations. We release both the code and the Emotionally Incongruent Synthetic Speech dataset (EMIS) to the community.

Why did Israel launch air strikes on Gaza? Will the US plan for Gaza fail? 'We survived the war, we may not survive the ceasefire' Spinning genocide: How is Israel using US PR firms to frame its Gaza war? Israel has contracted at least three public relations companies to bolster its image online and among the United States' Christian right, filings under the Foreign Agents Registration Act (FARA) show. According to US Department of Justice records, Israel's Ministry of Foreign Affairs hired the newly established Bridges Partners, the Christian PR agency Show Faith by Works, and the online consultancy Clock Tower X via the European Havas Media Group. Israel is acutely conscious of the need to control how its war, in which it has killed more than 68,000 Palestinians, is perceived by its allies and sponsors in the US .

The latest season of Netflix's The Witcher has landed - with one big difference. Former lead actor Henry Cavill has been replaced as main character Geralt of Rivia by Liam Hemsworth. The Australian has stepped in for the final two seasons of the fantasy show, based on a popular series of novels and video games. Previously, British actor Cavill had portrayed the title character, a monster hunter with supernatural abilities known as the White Wolf. When he announced he was passing the torch to Hemsworth in October 2022, describing him as a fantastic actor, not all fans agreed.

Breakthroughs, discoveries, and DIY tips sent every weekday. For the first time, a brown rat has been caught on camera actively hunting bats . The never-before-seen footage shows the rat grabbing a snack at hibernation sites in northern Germany. While it's undeniably impressive that rats can grab their supper mid-air, the new footage does not bode well for the bats. According to a study recently published in the journal, rat predation may cause enough damage to significantly threaten local bat populations .

Trump-Xi meeting: Who has the upper hand? Could Trump go for a third term? Is the US eyeing its next Latin American target? Why is Trump tearing down parts of the White House? United States President Donald Trump and his Chinese counterpart Xi Jinping have agreed to a trade truce under which the US will ease tariffs and Beijing will restart imports of US soya beans, delay the introduction of export restrictions on some of its rare earth metals and intensify efforts to curb illegal fentanyl trafficking.

Virgin Media O2 is set to become the first mobile operator to offer UK customers automatic connectivity via satellite in places without mobile signal. O2 Satellite will be an optional service due to launch in the first half of 2026. The firm has not yet revealed how much it will cost, but it will be an additional fee to pay each month. O2 has partnered with Elon Musk's satellite business Starlink to offer the service. Enabled smartphones will automatically switch to satellite coverage in parts of the UK where there is no terrestrial signal available - for example in rural areas.

Who is Rob Jetten, tipped to become youngest Dutch prime minister? Rob Jetten's achievement in dragging his socially liberal D66 party from fifth place to the top of Dutch politics in less than two years has been extraordinary. But politically, all the stars were perfectly aligned for the 38-year-old to do so. The result of Wednesday's election is too close to call, with Jetten vying with anti-Islam populist Geert Wilders for the most seats in parliament. No other political leader commanded as much screen time during the campaign as Jetten and his smile and cheerful message resonated with voters, while his rivals sometimes struggled.

Burnout is a psychological syndrome marked by emotional exhaustion, depersonalization, and reduced personal accomplishment, with a significant impact on individual well-being and organizational performance. This study proposes a machine learning approach to predict burnout risk using the HackerEarth Employee Burnout Challenge dataset. Three supervised algorithms were evaluated: nearest neighbors (KNN), random forest, and support vector machine (SVM), with model performance evaluated through 30-fold cross-validation using the determination coefficient (R2). Among the models tested, SVM achieved the highest predictive performance (R2 = 0.84) and was statistically superior to KNN and Random Forest based on paired $t$-tests. To ensure practical applicability, an interactive interface was developed using Streamlit, allowing non-technical users to input data and receive burnout risk predictions. The results highlight the potential of machine learning to support early detection of burnout and promote data-driven mental health strategies in organizational settings.

Despite their impressive performance, vision-language models (VLMs) still struggle on culturally situated inputs. To understand how VLMs process culturally grounded information, we study the presence of culture-sensitive neurons, i.e. neurons whose activations show preferential sensitivity to inputs associated with particular cultural contexts. We examine whether such neurons are important for culturally diverse visual question answering and where they are located. Using the CVQA benchmark, we identify neurons of culture selectivity and perform causal tests by deactivating the neurons flagged by different identification methods. Experiments on three VLMs across 25 cultural groups demonstrate the existence of neurons whose ablation disproportionately harms performance on questions about the corresponding cultures, while having minimal effects on others. Moreover, we propose a new margin-based selector - Contrastive Activation Selection (CAS), and show that it outperforms existing probability- and entropy-based methods in identifying culture-sensitive neurons. Finally, our layer-wise analyses reveals that such neurons tend to cluster in certain decoder layers. Overall, our findings shed new light on the internal organization of multimodal representations.