🤖 AI Summary
This work addresses the lack of systematic organization and unified evaluation frameworks for large language model (LLM) technologies. Methodologically, it proposes the first comprehensive, lifecycle-oriented technical framework covering pretraining, adaptation and fine-tuning, deployment and application, and capability evaluation. It integrates foundational components—including Transformer architectures, instruction tuning, prompt engineering, and reinforcement learning from human feedback (RLHF)—alongside multidimensional benchmarks (e.g., MMLU, BIG-Bench) to rigorously characterize emergent capabilities driven by scale transitions. Innovatively, it establishes a standardized evaluation methodology and an open-source resource taxonomy. The resulting authoritative survey delivers both theoretical insight and practical guidance: it systematically synthesizes technological evolution, core methodologies, open challenges, and high-quality resources, thereby providing a foundational reference for LLM research and real-world deployment.
📝 Abstract
Language is essentially a complex, intricate system of human expressions governed by grammatical rules. It poses a significant challenge to develop capable AI algorithms for comprehending and grasping a language. As a major approach, language modeling has been widely studied for language understanding and generation in the past two decades, evolving from statistical language models to neural language models. Recently, pre-trained language models (PLMs) have been proposed by pre-training Transformer models over large-scale corpora, showing strong capabilities in solving various NLP tasks. Since researchers have found that model scaling can lead to performance improvement, they further study the scaling effect by increasing the model size to an even larger size. Interestingly, when the parameter scale exceeds a certain level, these enlarged language models not only achieve a significant performance improvement but also show some special abilities that are not present in small-scale language models. To discriminate the difference in parameter scale, the research community has coined the term large language models (LLM) for the PLMs of significant size. Recently, the research on LLMs has been largely advanced by both academia and industry, and a remarkable progress is the launch of ChatGPT, which has attracted widespread attention from society. The technical evolution of LLMs has been making an important impact on the entire AI community, which would revolutionize the way how we develop and use AI algorithms. In this survey, we review the recent advances of LLMs by introducing the background, key findings, and mainstream techniques. In particular, we focus on four major aspects of LLMs, namely pre-training, adaptation tuning, utilization, and capacity evaluation. Besides, we also summarize the available resources for developing LLMs and discuss the remaining issues for future directions.
