Natural Language Processing

Researchers introduced AfriSUD, the first large-scale collection of syntactically annotated treebanks for nine African languages, using the Surface-Syntactic Universal Dependencies framework to capture typological featur…

A new expert-authored benchmark reveals that leading AI models achieve only 57-77% pass rates on multi-turn shopping conversations, with performance dropping 4-18 points as dialogues progress. The Shopping Reasoning Benc…

Researchers at arXiv have introduced a constrained semantic decompression task to test whether large language models can transform Persian proverbs into morally faithful narratives. Their study, using the new Proverb Ali…

Researchers have developed MARD (Mirror-Augmented Reasoning Distillation), a 7-billion-parameter AI system that predicts how drugs interact at the mechanism level, identifying specific enzymes or pathways involved rather…

Researchers have introduced MLUBench, a large-scale benchmark for evaluating lifelong unlearning in multimodal large language models (MLLMs), featuring 127 entities across 9 classes. The benchmark reveals that existing u…

Researchers at arXiv have introduced a framework for generating and validating question-answer datasets that test procedural reasoning in AI-supported learning systems. Comparing three strategies across 23 instructional …

A new study from arXiv (2606.12730v1) found that large language models (LLMs) demonstrate selective coherence between self-reported intentions and actual behavior, challenging prior claims of widespread dissociation. Res…

Researchers at an academic medical center developed a pre-response classifier that predicts the risk of a user rejecting an LLM-generated clinical response, achieving an AUROC of 0.719 over 4.5 months of user feedback. T…

Researchers have developed ToolSense, an open-source diagnostic framework that automatically generates three benchmarks to audit whether large language models truly understand their parametric tool knowledge. Testing on …

Researchers have identified a failure mode in multimodal AI systems called "Modal Isolation," where generated images and text reasoning fail to inform each other during complex tasks. The team proposes MoTiF, a two-stage…

Researchers introduced GRIP, a feedback-guided retrieval framework for multimodal in-context learning that selects examples based on their actual impact on model predictions rather than visual similarity. The system impr…

Researchers introduced Efficient Continual Alignment (ECA), a new exemplar-free incremental learning approach for open-ended image-to-text generation that adapts alignment modules within pre-trained vision-language model…

Researchers have identified three critical limitations in existing preference optimization methods for medical Large Vision-Language Models (LVLMs), including their inability to distinguish clinically critical tokens fro…

Researchers have introduced Rotary Value Embeddings (RoVE), a parameter-free modification to Rotary Position Embeddings (RoPE) that makes value pathways position-sensitive by rotating values simultaneously with keys. Thi…

A new study from researchers testing adapter composition in large language models found that geometry-aware merging strategies, which enforce orthogonality or directional independence in parameter updates, provide no con…

Researchers at arXiv introduced a dual-stance evaluation method to test whether activation steering on Llama-3-8B-Instruct reduces sycophancy without suppressing agreement with factually correct statements. The study fou…

Researchers released a multilingual vaccine benchmark called VaxEval containing 1,886 multiple-choice questions covering 14 vaccines in English, Spanish, and Chinese, drawing from sources including the World Health Organ…

Sophia NLU v1.2, an on-device, low-compute natural language engine built in Rust, now runs voice assistants without internet or a GPU, using only 160MB of RAM. The upgrade adds entity clarification, state persistence, an…

RAG (Retrieval-Augmented Generation) enables AI applications to retrieve relevant information from external data sources and use that context to generate accurate responses, overcoming the limitation of large language mo…

A developer replaced brittle regex and CSS selectors with a local LLM to extract product data from messy legacy HTML, achieving reliable results at a fraction of the cost of cloud-based models. The approach uses Ollama's…