LLM Agent Paper List

Report Abuse

Basic Information

This repository is a curated, continuously maintained survey and bibliography of research on large language model (LLM) based agents. It accompanies the paper "The Rise and Potential of Large Language Model Based Agents: A Survey" and collects must-read papers, categorized explanations, and updates about developments in agent design, capabilities, and evaluation. The README organizes material around a conceptual framework for LLM agents (brain, perception, action), practical applications (single agents, multi-agent systems, human-agent cooperation), agent societies, benchmarks, and training approaches. It highlights recent news and newly released works, lists project maintainers and contact information, and provides a citation for the survey. The repository is intended as a reference hub for researchers, students, and practitioners seeking an organized entry point into the literature on LLM-powered agents.

Links

Categorization

App Details

Features
The README provides an extensive table of contents and a structured taxonomy of topics including brain components (language, knowledge, memory, reasoning and planning), multimodal perception, action modalities (tool use and embodied control), applications, multi-agent coordination, and societal simulation. It lists many annotated papers with publication year, short notes, and code or project tags when available, and highlights benchmarks and datasets. The repo surfaces news about major releases, includes example figures and a BibTeX citation, names maintainers and contributors, and points to code and datasets referenced in the survey. The document is organized for browsing by topic and is open to contributions via pull requests, issues, and contact with the authors.
Use Cases
This collection helps readers quickly locate key papers, benchmarks, datasets, and implementation links relevant to LLM-based agents across subtopics such as memory systems, planning, tool use, embodied agents, multi-agent interaction, and evaluation. It reduces literature search time by grouping works by theme and by noting available code or project pages where present. The survey framing and ToC make it easier to understand research trends and open problems, and the included citation and maintainer contact support academic use and collaboration. Researchers can use it for literature reviews, students for learning pathways, and developers for finding prior implementations and benchmark resources.

Please fill the required fields*