This repository serves as a curated collection of research papers focused on Large Language Model (LLM)-powered agents, targeting researchers and practitioners in the AI field. It aims to benefit the community by organizing recent advancements in LLM-agent research published at top conferences, providing a valuable resource for understanding the evolving landscape of autonomous AI systems.

How It Works

The project organizes papers into thematic categories such as Agent Building, Profile, Memory, Planning, Action, Application, and Evaluation. Each entry includes a link to the paper and, where available, the associated code. Brief "TLDR" summaries highlight the core contribution of each paper, often explaining the novel approach or key insight, such as using role-playing for agent communication (CAMEL) or integrating fast and slow thinking modules for complex tasks (SwiftSage).

Quick Start & Requirements

This repository is a curated list of papers and does not have a direct installation or execution command. Users are directed to the survey "A Survey on Large Language Model based Autonomous Agents" and its associated repository for more comprehensive details.

Highlighted Details

• Organizes papers into key LLM-agent research areas: Building, Profile, Memory, Planning, Action, Application, and Evaluation.
• Provides direct links to research papers and, where available, their corresponding code repositories.
• Includes concise "TLDR" summaries for each paper, highlighting its main contribution and approach.
• Features papers from NIPS'23, indicating a focus on recent advancements.

Maintenance & Community

The project is maintained by Xueyang Feng, Lei Wang, and Chen Ma, with contributions noted for NIPS'23 papers. The README welcomes contributions and feedback from the research community.

Licensing & Compatibility

The licensing information for the repository itself is not explicitly stated in the README. The linked papers and code repositories will have their own respective licenses.

Limitations & Caveats

The repository is a curated list and does not provide executable code or a framework for building LLM agents. Its primary function is informational, pointing users to relevant research. Some entries mention "Awaiting publication," indicating that the research may not be fully accessible or finalized.