Unmasking falsehoods: A New Approach to AI Truthfulness
![A framework assesses language model reliability by extracting latent states from a frozen [latex]Qwen2.5-7B-Instruct[/latex] model and computing hallucination probabilities with neural network probes, enabling real-time detection of fabricated content as the system processes each token.](https://arxiv.org/html/2512.20949v1/x2.png)
Researchers have developed a novel method to detect when large language models are fabricating information, moving beyond simple accuracy metrics.
Science
Tracing the Outbreak: How Network AI Pinpoints Disease Origins
![A graph neural network addresses the challenge of identifying an epidemic’s origin by analyzing the network’s adjacency matrix and one-hot encoded node states at a given observation time [latex]t\_1[/latex], ultimately outputting a probability distribution across all nodes to pinpoint the most likely source of the outbreak.](https://arxiv.org/html/2512.20657v1/x1.png)
A new review examines the power of graph neural networks to rapidly and accurately identify the source of epidemics using network data.
Powering Up the Grid with AI Assistants
![The X-GridAgent system integrates four key features - [latex]F_1[/latex], [latex]F_2[/latex], [latex]F_3[/latex], and [latex]F_4[/latex] - to establish a robust framework for distributed grid navigation, acknowledging that even elegantly designed systems will inevitably encounter the unpredictable realities of production environments.](https://arxiv.org/html/2512.20789v1/Xgrid.png)
A new agentic AI system uses the power of large language models to streamline complex power grid analysis and automation.
Self-Optimizing Networks: The Rise of Agentic AI in 6G
A new framework leverages simulation and intelligent reflection to enable fully autonomous optimization of 6G radio access networks.
Beyond Calculation: Teaching Machines to Reason with Math
A new framework, AgentMath, dramatically improves the mathematical reasoning abilities of large language models by letting them actively use and learn from code execution.
Smarter Explanations, Better Decisions: An AI Framework for Actionable Insights
This research introduces an agentic approach to explainable AI that uses iterative refinement to improve the quality and usefulness of recommendations, particularly in complex domains like agriculture.
Sharper Focus: Training Transformers to Attend to What Matters
A new method refines attention mechanisms in Transformer models by dynamically identifying and correcting misleading attention patterns during training.
AI vs. the Data Scientist: When Code Isn’t Enough
New research reveals that current AI agents struggle to match human performance on complex data science tasks, particularly when domain expertise embedded in visual data is crucial.
Smarter Forecasts, Bigger Savings: Optimizing Demand with Dynamic Cost Control
A new approach to demand forecasting leverages node-level cost asymmetries and self-regulation to dramatically improve financial outcomes.
Reasoning with Data: A New Approach to Tabular Analysis
Researchers have developed a novel system that combines the power of large language models with reinforcement learning to dramatically improve performance on complex data reasoning tasks.