AI
A concentration result for multilayer feedforward neural networks
Researchers have made a breakthrough in understanding how multilayer feedforward neural networks behave. Specifically, they've found that if the weigh...
AI
AI
Agentic-SQL Revisited: Autonomy-Based Taxonomy and Empirical Benchmark Analysis for LLM Text-to-SQL
A new approach to evaluating Large Language Model (LLM) performance in text-to-SQL tasks is proposed. The authors argue that current evaluation method...
AI
TwinGridShield: Consequence-Aware Runtime Authorization for LLM Grid-Agent Actions
Researchers have developed a system called TwinGridShield to prevent large language models from making unauthorized changes to power grids. The system...
AI
Visible Reasoning and Indirect Prompt-Injection Monitorability Across English, Tamil, and Tanglish
Researchers studied how a language model responds to indirect prompts in three languages: English, Tamil, and a mix of both. They found that when the ...
AI
Large Language Model Assisted Operational Monitoring for Battery Energy Storage System Integrated Power Distribution Networks
Researchers have developed a framework for monitoring battery energy storage systems in power distribution networks using large language models. The s...
AI
Implementation of a Metacognition Framework for Self-Awareness and Self-Regulation in Ensembles of LLMs
Researchers have developed a metacognition framework to improve Large Language Models' (LLMs) ability to assess their own uncertainty and knowledge li...
AI
A survey of AI-generated voices and their detection
A recent arXiv paper surveys AI-generated voices and their detection methods. The authors highlight the rapid advancement of voice generation technolo...
AI
Does the Proof Prove It That Way? Faithful Formalization of Elements Proofs
Researchers have developed a method called Pistis to create formal proofs that accurately reflect the reasoning behind natural-language arguments. Thi...
AI
OTel: Building Domain-Specialized Telecom LLM Foundations for Intelligent Networks
Researchers have developed OTel, an open-source resource for building domain-specialized language models in the telecom industry. The project provides...
AI
Measuring Reward Hacking and Reasoning-Answer Decoupling Under Position-Confounded Optimization
Researchers have identified a problem in training language models where the model learns to exploit surface features rather than solving the task itse...
AI
Mental Model Management: An Operator-Based Framework for LLM Memory
Researchers have proposed a framework called Mental Model Management (3M) to help large language models like LLMs better organize and update their kno...
AI
Dynamic Multi-Byte Prediction With Hierarchical Language Models
Researchers have proposed a new approach to improve the efficiency of language models. They call it multi-byte prediction (MBP), which generates multi...