A Comprehensive Survey on Large Language Models: Architectures, Applications, and Ethical Considerations
Journal Title: Engineering and Technology Journal - Year 2025, Vol 10, Issue 04
Abstract
The rapid progress and revolutionary potential of large language models (LLMs) are examined in this survey, with particular attention paid to transformer-based designs and scaling strategies that power models such as GPT-4, Gemini, and LLaMA. Focusing on advancements in natural language comprehension, translation, text production, and information retrieval, this study examines various applications in the fields of healthcare, finance, education, and public administration. Along with addressing ethical issues such as algorithmic bias, disinformation, privacy threats, and environmental impacts, the study promotes robust auditing frameworks, consistent evaluation measures, and environmentally friendly practices. The study offers a comprehensive framework for future research and the responsible development of large language models (LLMs), outlining the complex supply chain and development processes with a focus on ethical integration, accountability, and transparency.
Authors and Affiliations
Chreesk Sabah M. Ali , Dr. Ibrahim Mahmood Ibrahim,
Keywords
Related Articles
- EP ID EP765133
- DOI 10.47191/etj/v10i04.26
- Views 11
- Downloads 0
Comparative Analysis of AI and Machine Learning Applications in Modern Database
Using artificial intelligence (AI) and machine learning (ML) in contemporary database systems is the topic of discussion in this article. It examines a variety of scientific publications that investigate the breakthrough...
KINETICS LEACHING STUDY OF URANIUM FROM ALGERIAN PHOSPHATE BY SULFURIC ACID
The kinetics of dissolution of an Algerian phosphate in sulfuric acid has been investigated in the presence of hydrogen peroxide (H2O2) as oxidant. The effect of different parameters such as leaching time, sulfuric acid...
CONSIDERATIONS PERTAINING TO EXERTED PIPING LOADS ON PRESSURE VESSEL NOZZLES
This technical note is intended to make the engineer involved in the design of pressure vessels with connected piping aware of the possible options that yield the most advantageous design from an economic perspective. Pr...
A DEEP-FRACTURE APPROACH TO EMBEDDED MULTILAYER NEURAL NETWORKS FOR HANDWRITTEN CHARACTER RECOGNITION
Low-cost microcontrollers, but also with reduced processing and memory capabilities, are highly appreciated devices for the development of specific tasks as embedded systems. The devices found in the market today provide...
Automating Legal Compliance and Contract Management: Advances in Data Analytics for Risk Assessment, Regulatory Adherence, and Negotiation Optimization
The integration of data analytics into legal compliance and contract management is transforming traditional processes by automating risk assessments, enhancing regulatory adherence, and optimizing negotiations. This pape...