Sanju Tiwari 

Date/Time: 05.06.2024, 12:00

Location: D2.2.094 

Abstract 

Large Language Models have presented an impressive performance and have become fundamental in real-world applications. These models are built upon vast amounts of text data and are trained to understand, generate, and manipulate human language with remarkable fluency. Large language models have applications in various fields, including content generation, language translation, sentiment analysis, and conversational agents. However, their development also raises ethical concerns regarding issues such as bias, misinformation, and data privacy, emphasizing the importance of responsible deployment and ongoing research in this rapidly evolving field. This talk delves into the multifaceted realm of LLMs, exploring the benchmarks, the biases inherent within them, and the diverse applications they fuel across various domains.  It will explore how LLMs are evaluated through benchmarks, discussing the challenges and opportunities in benchmark design and interpretation. Furthermore, it will also explore the biases present in LLMs, exploring their origins, impacts, and strategies for mitigation.