This material may not be published, broadcast, rewritten, or redistributed.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 .

Six new cookbooks bring stellar dishes--and cultures--from around the world into your kitchen. All products featured on WIRED are independently selected by our editors. However, we may receive compensation from retailers and/or from purchases of products through these links. Think about how difficult cooking from a cookbook from another culture was as little as 10 years ago. Once in a while, you could get your hands on a standout, but the food you could make with it could feel like a compromise with too many substitutions and ingredients you just couldn't find without great effort, or at all.

In this paper, we present a novel framework which accomplishes this by exploiting latent embedding spaces to define an objective function for an LLM in an iterative RL-driven process. As an example, consider the challenge of assisting content creators in generating valuable content within a recommender ecosystem (e.g., Y ouTube, Reddit, Spotify) [Boutilier et al., 2024].

Despite growing research on bias in large language models (LLMs), most work has focused on gender and race, with little attention to religious identity. This paper explores how religion is internally represented in LLMs and how it intersects with concepts of violence and geography. Using mechanistic interpretability and Sparse Autoencoders (SAEs) via the Neuronpedia API, we analyze latent feature activations across five models. We measure overlap between religion- and violence-related prompts and probe semantic patterns in activation contexts. While all five religions show comparable internal cohesion, Islam is more frequently linked to features associated with violent language. In contrast, geographic associations largely reflect real-world religious demographics, revealing how models embed both factual distributions and cultural stereotypes. These findings highlight the value of structural analysis in auditing not just outputs but also internal representations that shape model behavior.

Mogadishu, Somalia – Ending the United States' "forever wars" was a major slogan of Donald Trump's 2024 election campaign, during which he and many of his supporters spoke out against American resources and lives being put to waste in conflicts across the globe. But on February 1, a mere 10 days after being inaugurated for a second time, President Trump announced that the US had carried out air strikes targeting senior leadership of ISIL (ISIS) in Somalia. "These killers, who we found hiding in caves, threatened the United States," his post on X read. This marked Trump's first military action overseas, but it wouldn't be his last. In the time since, the US has provided weapons and support to Israel in its wars in Gaza and across the Middle East; it has launched strikes on Yemen; and even attacked Iran's nuclear facilities.

Historically, these systems included purely statistical models, but modern LLMs are deep artificial neural network s trained via machine learning . Once trained, an LLM may be implemented for various purposes, such as in chatbot s and personal assistants, or for translation, sentiment analysis and document review. 2 T he indisputably impressive performance of LLMs on a wide variety of task raises pressing questions about their capacities, and the mechanisms underlying those capacities . For instance, authors have grapple d with the questions of whether LLMs understand language (Bender & Koller, 2020; Mitchell & Krakauer, 2022) whether they possess concepts (Butlin, 2023) or to what extent they possess a theory of mind (Kosinski, 2024; Ullman 2023) . This paper focuses on the representational capacities of LLMs . D o LLMs rely on representations? If so, what do those representations represent? Much r esearch in AI -- for instance, studies using p robing classifiers (Belinkov, 2022), and methods for " e diting " models' representations (Hernandez et al., 202 4; Meng et al., 2022) -- assume s that a representational lens is appropriate . But a key question is whether LLMs can represent real world entities, or only "shallow" linguistic contents that don't reach into extra - linguistic reality (Butlin, 2021; Coelho Mollo & Millière, 2023; Yildirim & Paul, 2024) .

The fact that everyone with a social media account can create and share content, and the increasing public reliance on social media platforms as a news and information source bring about significant challenges such as misinformation, fake news, harmful content, etc. Although human content moderation may be useful to an extent and used by these platforms to flag posted materials, the use of AI models provides a more sustainable, scalable, and effective way to mitigate these harmful contents. However, low-resourced languages such as the Somali language face limitations in AI automation, including scarce annotated training datasets and lack of language models tailored to their unique linguistic characteristics. This paper presents part of our ongoing research work to bridge some of these gaps for the Somali language. In particular, we created two human-annotated social-media-sourced Somali datasets for two downstream applications, fake news \& toxicity classification, and developed a transformer-based monolingual Somali language model (named SomBERTa) -- the first of its kind to the best of our knowledge. SomBERTa is then fine-tuned and evaluated on toxic content, fake news and news topic classification datasets. Comparative evaluation analysis of the proposed model against related multilingual models (e.g., AfriBERTa, AfroXLMR, etc) demonstrated that SomBERTa consistently outperformed these comparators in both fake news and toxic content classification tasks while achieving the best average accuracy (87.99%) across all tasks. This research contributes to Somali NLP by offering a foundational language model and a replicable framework for other low-resource languages, promoting digital and AI inclusivity and linguistic diversity.

Large language models are increasingly being integrated into applications that shape the production and discovery of societal knowledge such as search, online education, and travel planning. As a result, language models will shape how people learn about, perceive and interact with global cultures making it important to consider whose knowledge systems and perspectives are represented in models. Recognizing this importance, increasingly work in Machine Learning and NLP has focused on evaluating gaps in global cultural representational distribution within outputs. However, more work is needed on developing benchmarks for cross-cultural impacts of language models that stem from a nuanced sociologically-aware conceptualization of cultural impact or harm. We join this line of work arguing for the need of metricizable evaluations of language technologies that interrogate and account for historical power inequities and differential impacts of representation on global cultures, particularly for cultures already under-represented in the digital corpora. We look at two concepts of erasure: omission: where cultures are not represented at all and simplification i.e. when cultural complexity is erased by presenting one-dimensional views of a rich culture. The former focuses on whether something is represented, and the latter on how it is represented. We focus our analysis on two task contexts with the potential to influence global cultural production. First, we probe representations that a language model produces about different places around the world when asked to describe these contexts. Second, we analyze the cultures represented in the travel recommendations produced by a set of language model applications. Our study shows ways in which the NLP community and application developers can begin to operationalize complex socio-cultural considerations into standard evaluations and benchmarks.

Retrieval-augmented generation (RAG) has demonstrated effectiveness in mitigating the hallucination problem of large language models (LLMs). However, the difficulty of aligning the retriever with the diverse LLMs' knowledge preferences inevitably poses an inevitable challenge in developing a reliable RAG system. To address this issue, we propose DPA-RAG, a universal framework designed to align diverse knowledge preferences within RAG systems. Specifically, we initially introduce a preference knowledge construction pipline and incorporate five novel query augmentation strategies to alleviate preference data scarcity. Based on preference data, DPA-RAG accomplishes both external and internal preference alignment: 1) It jointly integrate pair-wise, point-wise, and contrastive preference alignment abilities into the reranker, achieving external preference alignment among RAG components. 2) It further introduces a pre-aligned stage before vanilla Supervised Fine-tuning (SFT), enabling LLMs to implicitly capture knowledge aligned with their reasoning preferences, achieving LLMs' internal alignment. Experimental results across four knowledge-intensive QA datasets demonstrate that DPA-RAG outperforms all baselines and seamlessly integrates both black-box and open-sourced LLM readers. Further qualitative analysis and discussions also provide empirical guidance for achieving reliable RAG systems. Our code is publicly available at https://github.com/dongguanting/DPA-RAG.