History of Large Language Models
摘要
This chapter explores the historical development of artificial intelligence (AI) and natural language processing (NLP), focusing on the evolution of language modeling. We begin by outlining the foundations of AI, from symbolic approaches to the emergence of sub-symbolic approaches such as machine learning (ML) and deep learning (DL). We examine NLP advancements, distinguishing between natural language understanding (NLU) and natural language generation (NLG) and highlighting the innovative role of the transformers architectures. The chapter then delves into the progression of language models (LMs), from early statistical LMs to neural LMs and the modern era of large language models (LLMs). We present a historical timeline of key LLMs—including GPT, PaLM, and LLaMA—alongside their technological milestones, shedding light on their increasing scale, capabilities, and impact.