This repository serves as a curated collection of survey papers and research on Large Language Models (LLMs). It aims to provide a comprehensive overview of the rapidly evolving LLM landscape, covering key areas such as instruction tuning, human alignment, LLM agents, hallucination, multimodality, and domain-specific applications. The project is intended for researchers and practitioners seeking to stay abreast of LLM advancements and contribute to the collective knowledge base.

How It Works

The repository organizes survey papers by topic, creating a structured and easily navigable resource. Each entry typically includes a link to the paper and, where available, a project page. This approach facilitates quick access to relevant research, enabling users to efficiently explore specific subfields within LLM development and application.

Quick Start & Requirements

This is a curated list of research papers and does not involve code execution or installation. Users can directly access the listed papers via provided links.

Highlighted Details

• Extensive categorization of LLM research, from foundational concepts to specialized applications like LLMs in Health, Finance, Law, and Robotics.
• Covers emerging and critical LLM challenges including hallucination mitigation, efficiency, security, privacy, and detection of LLM-generated content.
• Includes surveys on advanced LLM capabilities such as Chain-of-Thought prompting, Retrieval-Augmented Generation (RAG), and LLM-based agents.
• Features research on multimodal LLMs, encompassing vision-language, audio, and code-related LLM applications.

Maintenance & Community

The repository encourages community contributions via pull requests to update paper information, fostering a collaborative environment for tracking LLM research.

Licensing & Compatibility

The repository itself is a collection of links to external research papers. The licensing and compatibility of the individual papers are determined by their respective publishers or authors.

Limitations & Caveats

As a survey repository, it relies on community contributions for updates and may not always reflect the absolute latest research immediately upon publication. The accuracy and completeness of information are dependent on the quality of submitted pull requests.