explore | Grilled Cheese

New arXiv work analyzes JEV and three open KEV direct-decision models, finding they compress ordinal scales across 36 datasets, using only 67-76% of effective gold support versus 87-102% on nominal tasks. Randomizing candidate…

#OpenSourceAI #LLMJudges #AIBias #NLP
https://arxiv.org/abs/2609.38827

Die Sprache der Maschinen meistern! 💡💬 NLP ist mega spannend! Starte dein Wissen hier: 👇 https://www.datenbanken-verstehen.de/lexikon/natural-language-processing #NLP #KI 🧠

A new arXiv study leverages WhisperX and Silero VAD with refined timestamping to automatically estimate the Verbal Fluency Index for cognitive monitoring in motor neuron disease. The approach using clinically inspired features…

#AI #NLP #DataScience #HealthTech
https://arxiv.org/abs/2609.38203

BARRAC adapts an English aspect-based sentiment analysis framework to Arabic dialect classification, replacing consumer-review attribute pools with Arabic linguistic devices and two-stage training. It achieves a mean macro-F1 of 63.93% on five…

#AI #NLP #ArabicNLP
https://arxiv.org/abs/2609.38820

A new benchmark, CineSubBench, tests LLMs on long-context film understanding using 1,012 films and 6,072 multilingual subtitle tracks totaling 8.13M timestamped entries. It targets narrative integration, causal reasoning, and culturally…

#LLMs #NLP #Benchmarks
https://arxiv.org/abs/2609.36218

Grateful to see our research on AI, NLP, and cardiovascular imaging selected by Elsevier for circulation in connection with National Heart Day. A beautiful continuation of the journey from clinical questions to data to published science.

doi.org/10.1016/j.ah...

#Cardiology #MedicalAI #NLP

New arXiv work quantifies the hidden compression cost of tokeniser rules, showing pre-tokenisation boundaries raise optimal token counts by 28–37% on English Wikipedia. Byte pair encoding sits close to constrained bounds yet further from…

#AI #NLP #Tokenisation
https://arxiv.org/abs/2609.35869

Teachers and Department Leaders! Opening questions for student conversation. What was the hardest part of today? Did anything surprise you today?

Purchase here:nbsp;https://tinyurl.com/3td7nc5k

#teachers #education #coaching #NeuroLinguisticProgramming #NLP #neurodiversity #studentsuccess

New paper REALHOP audits multi-hop reasoning benchmarks and finds Behavioral Necessity Rates of only 16.6% to 48.9%, meaning models often answer correctly without relying on the intended intermediate evidence. Suggests current…

#OpenSourceAI #AIResearch #NLP
https://arxiv.org/abs/2609.36984

CORTEX introduces a learning-dynamics framework that learns internal modularization within dense language models by partitioning trainable matrices into parameter groups based on gradient similarity. This approach lets models share…

#AI #LLMs #MachineLearning #NLP
https://arxiv.org/abs/2609.34449

LLMs appear to store substantial lexical knowledge, but new research examines how much of it they actually use during reasoning tasks like textual entailment. The study tests whether LLMs can supply the lexical relations needed…

#AIResearch #NLP #Entailment #LLMs
https://arxiv.org/abs/2609.32491

Here are the worst psychologist jokes ever, although a few are truly funny. The list is free and complete at worlds-worst-website.com/worst-psycho... ( #psychologist, #psychology, #psychiatrist, #psychiatry, #NLP, #joke, #humor, #psychologistJoke, #therapy, #counselor, #marriageCounselor)

OpenFC post-trains Qwen3-8B as a unified verification policy for open-search fact checking, learning reasoning, evidence acquisition, and stopping as a single next-action task via stepwise-calibrated cold start and supervised…

#OpenSource #AI #NLP #MachineLearning
https://arxiv.org/abs/2609.33579

A new arXiv paper studies the theory of hard, discrete prompts for large language models, showing that deciding whether a hard prompt exists for a transformer to solve a task is NP-complete and finding an optimal one is…

#PromptEngineering #LLM #AIResearch #NLP
https://arxiv.org/abs/2609.32302

Robots are no longer limited to repetitive motions behind safety cages. 🤖📦

This system can identify, grip, lift, move and stack boxes with impressive precision — turning physical AI into a practical warehouse tool.

#Robotics #Automation #PhysicalAI #FutureOfWork #AI #IoT #5G #NLP #Tech

Machine learning and rule-based models classified portal messages accurately, with best methods varying by message type; this could support faster triage. #MachineLearning #PatientPortals #NLP doi.org/10.1016/j.ij...