Thought, the LLM can verbalize that the unknown city is Paris and use this fact to answer downstream questions.

Citations from LLM-based RAG systems are supposed to simplify response verification. However, this does not hold for citation failure, when a model generates a helpful response, but fails to cite complete evidence. In contrast to previous work, we propose to disentangle this from response failure, where the response itself is flawed, and citing complete evidence is impossible. To address citation failure, this work follows a two-step approach: (1) We study when citation failure occurs and (2) how it can be mitigated. For step 1, we extend prior work by investigating how the relation between response and evidence affects citation quality. We introduce CITECONTROL, a benchmark that systematically varies this relation to analyze failure modes. Experiments show that failures increase with relational complexity and suggest that combining citation methods could improve performance, motivating step 2. To improve LLM citation efficiently, we propose CITENTION, a framework integrating generative, attention-based, and retrieval-based methods. Results demonstrate substantial citation improvements on CITECONTROL and in transfer settings. We make our data and code publicly available.

Thought, the LLM can verbalize that the unknown city is Paris and use this fact to answer downstream questions.

The secret cult caves of polyamorous Mormon'prophet' with 85 wives are seen for first time Florida's housing market is tanking but the birthplace of Southern rock keeps its groove and defies the crash My war with Harry & Meghan, by PIERS MORGAN: What really happened, their absurd accusations, the brutal truth about post-royal life... and how I believe their royal racism lies helped kill off woke But experts warn the huge benefits come with risks... here's what it means for YOU I hung ICE agent effigies from the gallows in my yard. MAGA had a huge meltdown. They're going to lose their minds when they see what else I've done Vile Chicago woman filmed rubbing dog poop on Cybertruck emblazoned with Donald Trump's signature Taylor, your album should be'Life of a Callgirl'. KENNEDY's appalled take on Swift's new record... and its ultra-vivid sex shout outs for Travis the Sasquatch Fate of the four Scottish crime lords who terrorised Dubai: Gangsters thought they were'untouchable' after spree of executions and firebombings. Now we reveal hellhole jail, inhumane'toilet paper' punishment... and where they are now Olympic gold medalist forced to put Louisiana home up for sale as she'can't make a living' months after filing for divorce Tycoon who is cousin of former President George W. Bush expected to launch run for Maine governor Israel prepares to implement'first stage' of Trump's Gaza peace plan Cassie Ventura's attorney responds to Diddy sentencing as she's hailed by judge who jailed vile rapper The truth about Keith Urban's guitarist'other woman' Maggie Baugh revealed amid Nicole Kidman divorce How I look like this at 62. I've lost 5 stone fast, 20 years off my biological age and wear size 8... without weight-loss jabs.

ASR has achieved remarkable global progress, yet African low-resource languages remain rigorously underrepresented, producing barriers to digital inclusion across the continent with more than +2000 languages. This systematic literature review (SLR) explores research on ASR for African languages with a focus on datasets, models and training methods, evaluation techniques, challenges, and recommends future directions. We employ the PRISMA 2020 procedures and search DBLP, ACM Digital Library, Google Scholar, Semantic Scholar, and arXiv for studies published between January 2020 and July 2025. We include studies related to ASR datasets, models or metrics for African languages, while excluding non-African, duplicates, and low-quality studies (score <3/5). We screen 71 out of 2,062 records and we record a total of 74 datasets across 111 languages, encompassing approximately 11,206 hours of speech. Fewer than 15% of research provided reproducible materials, and dataset licensing is not clear. Self-supervised and transfer learning techniques are promising, but are hindered by limited pre-training data, inadequate coverage of dialects, and the availability of resources. Most of the researchers use Word Error Rate (WER), with very minimal use of linguistically informed scores such as Character Error Rate (CER) or Diacritic Error Rate (DER), and thus with limited application in tonal and morphologically rich languages. The existing evidence on ASR systems is inconsistent, hindered by issues like dataset availability, poor annotations, licensing uncertainties, and limited benchmarking. Nevertheless, the rise of community-driven initiatives and methodological advancements indicates a pathway for improvement. Sustainable development for this area will also include stakeholder partnership, creation of ethically well-balanced datasets, use of lightweight modelling techniques, and active benchmarking.

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.

Large Language Models (LLMs) have achieved high accuracy on complex commonsense and mathematical problems that involve the composition of multiple reasoning steps. However, current compositional benchmarks testing these skills tend to focus on either commonsense or math reasoning, whereas LLM agents solving real-world tasks would require a combination of both. In this work, we introduce an Agentic Commonsense and Math benchmark (AgentCoMa), where each compositional task requires a commonsense reasoning step and a math reasoning step. We test it on 61 LLMs of different sizes, model families, and training strategies. We find that LLMs can usually solve both steps in isolation, yet their accuracy drops by ~30% on average when the two are combined. This is a substantially greater performance gap than the one we observe in prior compositional benchmarks that combine multiple steps of the same reasoning type. In contrast, non-expert human annotators can solve the compositional questions and the individual steps in AgentCoMa with similarly high accuracy. Furthermore, we conduct a series of interpretability studies to better understand the performance gap, examining neuron patterns, attention maps and membership inference. Our work underscores a substantial degree of model brittleness in the context of mixed-type compositional reasoning and offers a test bed for future improvement.

In the quest of accurate vertex classification, we introduce GRAVITY (Graph-based Representation leArning via Vertices Interaction TopologY), a framework inspired by physical systems where objects self-organize under attractive forces. GRAVITY models each vertex as exerting influence through learned interactions shaped by structural proximity and attribute similarity. These interactions induce a latent potential field in which vertices move toward energy efficient positions, coalescing around class-consistent attractors and distancing themselves from unrelated groups. Unlike traditional message-passing schemes with static neighborhoods, GRAVITY adaptively modulates the receptive field of each vertex based on a learned force function, enabling dynamic aggregation driven by context. This field-driven organization sharpens class boundaries and promotes semantic coherence within latent clusters. Experiments on real-world benchmarks show that GRAVITY yields competitive embeddings, excelling in both transductive and inductive vertex classification tasks.

We propose LingBench++, a linguistically-informed benchmark and reasoning framework designed to evaluate large language models (LLMs) on complex linguistic tasks inspired by the International Linguistics Olympiad (IOL). Unlike prior benchmarks that focus solely on final answer accuracy, LingBench++ provides structured reasoning traces, stepwise evaluation protocols, and rich typological metadata across over 90 low-resource and cross-cultural languages. We further develop a multi-agent architecture integrating grammatical knowledge retrieval, tool-augmented reasoning, and deliberate hypothesis testing. Through systematic comparisons of baseline and our proposed agentic models, we demonstrate that models equipped with external knowledge sources and iterative reasoning outperform single-pass approaches in both accuracy and interpretability. LingBench++ offers a comprehensive foundation for advancing linguistically grounded, culturally informed, and cognitively plausible reasoning in LLMs.

This study explores the use of Federated Learning (FL) for tuberculosis (TB) diagnosis using chest X-rays in low-resource settings across Africa. FL allows hospitals to collaboratively train AI models without sharing raw patient data, addressing privacy concerns and data scarcity that hinder traditional centralized models. The research involved hospitals and research centers in eight African countries. Most sites used local datasets, while Ghana and The Gambia used public ones. The study compared locally trained models with a federated model built across all institutions to evaluate FL's real-world feasibility. Despite its promise, implementing FL in sub-Saharan Africa faces challenges such as poor infrastructure, unreliable internet, limited digital literacy, and weak AI regulations. Some institutions were also reluctant to share model updates due to data control concerns. In conclusion, FL shows strong potential for enabling AI-driven healthcare in underserved regions, but broader adoption will require improvements in infrastructure, education, and regulatory support.