Awesome AI Agents › Multi-Agent Surveys

AGI-Edgerunners/LLM-Agents-Papers

⭐ 2344 Python repository created 2023-05-31

The AGI-Edgerunners/LLM-Agents-Papers repository is a comprehensive and meticulously curated collection of academic papers focused on large language model (LLM) based agents. This repository serves as an extensive resource for researchers, developers, and enthusiasts interested in the latest advancements, methodologies, applications, and challenges related to LLM agents. It categorizes papers into various thematic areas such as surveys, techniques for enhancement, interaction modes, applications across diverse fields, automation, training methods, scaling strategies, stability concerns, infrastructure, and other relevant topics. The repository includes surveys that provide broad overviews and evaluations of LLM-based agents, covering multi-turn conversations, question answering, multi-agent systems, and domain-specific applications like medicine, finance, and software engineering. Techniques for enhancement focus on improving agent capabilities through planning, memory mechanisms, feedback and reflection, retrieval-augmented generation (RAG), and search strategies. Interaction categories explore role-playing, conversation, game playing, human-agent interaction, tool usage, and simulation, highlighting the diverse ways LLM agents can engage with users and environments. Applications span numerous disciplines including math, chemistry, biology, physics, geography, art, medicine, finance, software engineering, and research, demonstrating the versatility of LLM agents. Automation and training sections discuss workflow automation, automatic evaluation, fine-tuning, reinforcement learning, and direct preference optimization (DPO). Scaling addresses frameworks for single-agent and multi-agent systems, while stability covers safety, bias, and hallucination issues. Infrastructure includes benchmarks, evaluation platforms, and datasets. Additionally, the repository recommends other valuable paper lists for comprehensive reading, making it a central hub for academic literature on LLM agents. This resource is invaluable for anyone looking to stay updated on the state-of-the-art research and development in the field of LLM-based autonomous agents.

https://github.com/AGI-Edgerunners/LLM-Agents-Papers

agentsapplicationsartautomatic-evaluationautomationbenchmarkbiasbiologychemistryconversationdatasetdpoenhancement-techniquesenvironmentevaluationfeedbackfinancefine-tuninggame-playinggeographyhallucinationhuman-agent-interactioninfrastructureinteractionlarge-language-modelsllm-agentllm-agentsmathmedicinememory-mechanismmulti-agent-systempaper-listphysicsplanningplatformragreflectionreinforcement-learningresearchretrieval-augmented-generationrole-playingsafetyscalingsearchsimulationsingle-agent-frameworksoftware-engineeringstabilitysurveystool-usagetrainingworkflow