large language models

$S^3$-Bench: Evaluating Speech Interaction Models as Scientific Voice Assistants

S$^3$-Bench is introduced, a systematic evaluation framework covering 10 major disciplines, consisting of a Knowledge set for speech question-answering and a Dialogue set for multi-turn progressive interactions with simulated user agents that systematically characterize the common challenges and performance tradeoffs o...

Do LLMs Make More Mistakes If They Do Not Believe the Input Data?

How faithfulness of LLMs to provided context depends on how plausible they perceive the context to be (context-memory conflict) is analyzed, which makes use of the increased difficulty of non-English and low-resource language text generation and input data based on local knowledge.

Vague2Detect: Handling Ambiguous Prompts in Knowledge-Based Open-World Detection

This work proposes Vague2Detect, a hybrid pipeline in which a fine-tuned Sentence-BERT retrieves candidates from a structured household Knowledge Base, and YOLO-World verifies their presence in the image, and improves performance to 61% VPSR with high precision, and up to 85% when augmented with GPT fallback.

Emergency Department Revisit Quality Review Screening: Exploring Human Decision-Making and Artificial Intelligence Support

An exploratory, retrospective study of randomly selected ED visits to a multihospital health system having an ED revisit within 1-14 days to the same health system finds that the KGA may enable enhancing the scope and yield of screening without substantially increasing reviewer workload.

Beyond One-Size-Fits-All: Sample-Adaptive Strategy Routing for Vision Token Pruning in MLLMs

VIsion Pruning Router is proposed, a lightweight VIsion Pruning Router that adaptively selects the pruning strategy predicted to be best suited to each input at a specified pruning level that proves effective across various MLLM backbones and yields consistent gains on unseen benchmarks, highlighting the potential of s...

Active Adaptation, Not Static Defense: Temporal Dynamics of Preventative Steering in Adversarial Fine-Tuning

It is shown that preserving or reinjecting the weight offset fails to maintain protection, indicating that preventative steering relies on active adaptation rather than a static defense, and proposes Progressive Intensity Scheduling (PIS), which improves safety robustness over static-strength steering while reducing ha...

Modality-Decoupled Federated Learning for Privacy-Preserving Embodied Intelligence in 6G

FedMVLA is proposed, a modality-decoupled FL framework for privacy-preserving embodied intelligence in 6G networks that incorporates three mechanisms: modality-aware federated aggregation (MAFA), modality-aware privacy allocation (MAPA), and modality-aware communication compression (MACO), together with a modality-slic...

Beyond Surface Imitation: Contrastive Modeling for Reasoning Path Alignment in Multimodal In-Context Learning

A new multimodal ICL framework is proposed that combines contrastive demonstration modeling with the self-refinement capability of MLLMs and consistently improves MLLM performance, with particularly notable gains on visual question answering (VQA).

OntologyAligner: Ontology-Aligned Retrieval and Hierarchy-Guided Large Language Model Reranking for Biomedical Ontology Normalization

OntologyAligner is presented, a three-stage framework that combines ontology-aligned retrieval, large language model candidate reranking, and selective hierarchy-guided refinement that offers a generalizable framework for accurate mapping of biomedical text to structured ontology concepts.

Multi-Agent Agentic Graph Learning via Structural Signatures

A multi-agent agentic graph learning framework that partitions the graph into communities and assigns an independent agent to each community for region-specific specialization and shows that MAAGL outperforms SOTA AGL methods.

Can large language models replace human surveys in geospatial studies? A case of vague cognitive regions

Do large language models share the same culture? A psychometric framework for quantifying cross-model response divergence (Cross-Model Response Divergence, CMRD)