LLMs
Latest breakthroughs in Large Language Models
Articles
Closed-Loop LLM Co-Pilots for Digital Agriculture⭐8
LLM co-pilots autonomously optimize crop production, cutting energy use by 68% and cycle times by 35% in closed-loop digital agriculture trials.
Unified Hallucination Fuzzing for Multimodal Large Language Models⭐10
Unified fuzzing framework reveals severe hallucination risks in MLLMs, exposing hidden performance degradation and alignment trade-offs.
Towards an Argumentative Foundation for Evaluative AI⭐10
This paper proposes computational argumentation as a formal foundation for evaluative AI, enabling explainable, contestable decisions in critical domains.
TEXAS: Task-Expert-Aware Supervision for Downstream⭐9
TEXAS method improves MoE LLM adaptation by identifying task-relevant experts and optimizing token supervision, boosting performance across benchmarks.
Towards Multi-Label Graph Foundation Models: From Single-Vector⭐7
This paper introduces MSB-GFM, a multi-semantic basis graph foundation model for cross-domain multi-label node classification, addressing semantic
Simulator-Grounded Large Language Models for Industrial Causal⭐8
Compare three methods for grounding LLMs in a wastewater simulator, achieving up to 99.5% causal QA accuracy and robust cross-plant transfer.
Agentic Nesting: A New Methodology for Enterprise Application⭐7
Agentic Nesting converts legacy enterprise apps into AI agents within hierarchical structures, enabling natural-language integration and autonomous orchestratio...
Transfer Learning for Named Entity Recognition of Classical⭐8
Team uOttawa wins EvaLatin 2026 NER shared task using LLM prompting for Classical Latin, achieving top scores on coarse and fine-grained entity recognition.
ISEE: Interactive Semantic Enrichment for Database Fields⭐8
ISEE: Interactive Semantic Enrichment for Database Fields improves data clarity, reduces cognitive load, and boosts LLM agent performance in downstream tasks.
TabletCraft: Enabling Bidirectional Akkadian Translation and⭐7
TabletCraft enables bidirectional Akkadian-English translation and cuneiform rendering, bridging a 4,000-year cultural gap for broader public access.