Towards Understanding Large Language Models for Multilingual Semantic Encoding
Session Number
Project ID: CMPS 23
Advisor(s)
Ermin Wei, Northwestern University
Discipline
Computer Science
Start Date
17-4-2024 9:40 AM
End Date
17-4-2024 9:55 AM
Abstract
Natural Language Processing (NLP) has witnessed significant advancements with the emergence of large language models (LLM) capable of understanding and generating human-like text. However, there remains a critical need to explore and understand their efficiency and effectiveness, especially in processing languages beyond English. This study aims to evaluate the efficiency of various large language models in capturing semantic meaning across English, German, and Spanish sentences. Principal Component Analysis (PCA) is utilized to identify important weights for understanding semantics. Through further experimentation with various sentence structures, we aim to identify factors contributing to the effectiveness of certain models. By pinpointing the strengths and weaknesses of different models, we aim to advance NLP research for multilingual applications.
Towards Understanding Large Language Models for Multilingual Semantic Encoding
Natural Language Processing (NLP) has witnessed significant advancements with the emergence of large language models (LLM) capable of understanding and generating human-like text. However, there remains a critical need to explore and understand their efficiency and effectiveness, especially in processing languages beyond English. This study aims to evaluate the efficiency of various large language models in capturing semantic meaning across English, German, and Spanish sentences. Principal Component Analysis (PCA) is utilized to identify important weights for understanding semantics. Through further experimentation with various sentence structures, we aim to identify factors contributing to the effectiveness of certain models. By pinpointing the strengths and weaknesses of different models, we aim to advance NLP research for multilingual applications.