Natural Language Processing

Researchers at the University of Oxford and other institutions found that large language models (LLMs) develop false-belief task (FBT) performance late in pretraining, dependent on model size and training volume, and imp…

Researchers introduced a French OSCE dialogue dataset with 240 student-patient interactions and developed a controllable LLM-based pipeline to generate synthetic OSCE dialogues for clinical training. The system includes …

Researchers at arXiv show that static per-layer Fibonacci spacing in sparse attention outperforms learned dilation and fixed schedules across 21 language models, achieving better perplexity and extrapolating to four time…

Researchers propose grain calibration to validate whether large language models measure theoretical constructs correctly, not just reliably. The method decomposes constructs into clause-level components and tests each wi…

Researchers evaluated the phonological perception of deep learning models for sign language recognition, finding that pose-based models are sensitive to handshape contrasts while pixel-based models better capture locatio…

Researchers introduced SEATauBench, the first agent-focused evaluation framework for Southeast Asian sovereign AI, adapting TauBench to five languages. Testing across three models revealed that English agent capabilities…

Researchers introduced 5ting, a system for the SemEval-2026 Task 8 (MTRAGEval) that combines BGE-M3 dense retrieval, FAISS indexing, dual-query merged retrieval, and LLM-based reranking with role-separated generation con…

Researchers developed a structure-preserving Marathi-to-English document translation framework using a multi-stage LLM pipeline, integrating layout-aware OCR and HTML reconstruction to maintain formatting and domain term…

Researchers at arXiv found that using large language models (LLMs) like GPT-5.2 to label training data for entity matching can reduce manual labeling effort by hundreds of hours, with student models performing within two…

A new audit of over twenty African NLP corpus families reveals widespread license incompatibilities, including CC-BY-SA and CC-BY-NC datasets that cannot be legally combined and NoDerivs clauses that prohibit tokenisatio…

Researchers introduced SciDraw-Bench, a benchmark for evaluating scientific figure generation by text-to-image and multimodal models, covering 32 tasks across eight figure types and ten disciplines. In a pilot study, a d…

Researchers introduced SCSuff, an information-theoretic metric to evaluate whether free-text explanations from large language models (LLMs) are sufficient to explain their outputs. Testing on LLMs showed explanations are…

Researchers used frozen medical large language model (LLM) representations to predict primary ICD categories from both clinical narratives and structured EHR data, achieving 87.69% strict accuracy on MIMIC-IV. The multim…

A new knowledge representation architecture called Synapedia defines concepts using events sourced from Wikipedia, addressing a structural flaw in existing systems like WordNet and Wikidata. The system uses four layers a…

Andrej Karpathy declared prompt engineering dead in a June 25, 2025 tweet, introducing context engineering as its replacement. The shift redefines how developers interact with AI models, emphasizing structured context ov…

A developer built an AI tool that uses embeddings and retrieval-augmented generation to match incident alerts with runbook fixes by meaning rather than keyword overlap, closing the gap between alert messages and document…

A new survey paper on arXiv provides a comprehensive taxonomy of memory systems for AI agents, categorizing memory into forms (token-level, parametric, latent), functions (factual, experiential, working), and dynamics. T…

A developer proposes using AI agents to activate dormant 'weak ties' in professional networks, addressing the gap between having many connections and being able to identify collaboration opportunities. The approach lever…

Opportunity Skill, a new AI agent skill, automatically extracts and maintains a real-time professional interface definition from a user's daily work, eliminating the need for manual profile updates. The skill analyzes re…

Australian universities are using AI detection software to accuse students of cheating, with internal documents from Australian Catholic University showing nearly 6,000 misconduct referrals in a year, 90% involving suspe…