LeiWangetal. ASurveyonLargeLanguageModelbasedAutonomousAgents 35
basedrecommendationsystem. In: InternationalCon- code generation via chatgpt. arXiv preprint
ferenceonMachineLearning. 2019,1052–1061 arXiv:2304.07590,2023
12. ShinnN,CassanoF,GopinathA,NarasimhanK,YaoS. 25. SafdariM,Serapio-GarcíaG,CrepyC,FitzS,Romero
Reflexion: Languageagentswithverbalreinforcement P, Sun L, Abdulhai M, Faust A, Mataric´ M. Person-
learning. AdvancesinNeuralInformationProcessing ality traits in large language models. arXiv preprint
Systems,2024,36 arXiv:2307.00184,2023
13. ShenY,SongK,TanX,LiD,LuW,ZhuangY. Hug- 26. JohnsonJA. Measuringthirtyfacetsofthefivefactor
ginggpt: Solvingaitaskswithchatgptanditsfriends modelwitha120-itempublicdomaininventory: De-
inhuggingface. AdvancesinNeuralInformationPro- velopmentoftheipip-neo-120. Journalofresearchin
cessingSystems,2024,36 personality,2014,51: 78–89
14. Qin Y, Liang S, Ye Y, Zhu K, Yan L, Lu Y, Lin Y, 27. JohnOP,DonahueEM,KentleRL.Bigfiveinventory.
CongX,TangX,QianB,others. Toolllm: Facilitating JournalofPersonalityandSocialPsychology,1991
large language models to master 16000+ real-world 28. Deshpande A, Murahari V, Rajpurohit T, Kalyan
apis. arXivpreprintarXiv:2307.16789,2023 A, Narasimhan K. Toxicity in chatgpt: Analyzing
15. SchickT,Dwivedi-YuJ,DessìR,RaileanuR,Lomeli persona-assigned language models. arXiv preprint
M,HambroE,ZettlemoyerL,CanceddaN,ScialomT. arXiv:2304.05335,2023
Toolformer: Languagemodelscanteachthemselvesto 29. ArgyleLP,BusbyEC,FuldaN,GublerJR,RyttingC,
usetools. AdvancesinNeuralInformationProcessing WingateD. Outofone,many: Usinglanguagemodels
Systems,2024,36 tosimulatehumansamples. PoliticalAnalysis,2023,
16. ZhuX,ChenY,TianH,TaoC,SuW,YangC,Huang 31(3): 337–351
G,LiB,LuL,WangX,others. Ghostintheminecraft: 30. FischerKA. Reflectivelinguisticprogramming(rlp):
Generallycapableagentsforopen-worldenviroments A stepping stone in socially-aware agi (socialagi).
vialargelanguagemodelswithtext-basedknowledge arXivpreprintarXiv:2305.12647,2023
andmemory. arXivpreprintarXiv:2305.17144,2023 31. RanaK,HavilandJ,GargS,Abou-ChakraJ,ReidI,
17. SclarM,KumarS,WestP,SuhrA,ChoiY,Tsvetkov SuenderhaufN. Sayplan: Groundinglargelanguage
Y. Mindinglanguagemodels’(lackof)theoryofmind: modelsusing3dscenegraphsforscalablerobottask
Aplug-and-playmulti-characterbelieftracker. arXiv planning. In: 7thAnnualConferenceonRobotLearn-
preprintarXiv:2306.00924,2023 ing. 2023
18. QianC,CongX,YangC,ChenW,SuY,XuJ,LiuZ, 32. ZhuA,MartinL,HeadA,Callison-BurchC. Calypso:
SunM. Communicativeagentsforsoftwaredevelop- Llmsasdungeonmaster’sassistants. In: Proceedings
ment. arXivpreprintarXiv:2307.07924,2023 oftheAAAIConferenceonArtificialIntelligenceand
19. al. e C. Agentverse. https://github.com/ InteractiveDigitalEntertainment. 2023,380–390
OpenBMB/AgentVerse,2023 33. WangZ,CaiS,ChenG,LiuA,MaX,LiangY. De-
20. Park J S, O’Brien J, Cai C J, Morris M R, Liang P, scribe, explain, planandselect: Interactiveplanning
BernsteinMS. Generativeagents: Interactivesimu- withlargelanguagemodelsenablesopen-worldmulti-
lacraofhumanbehavior. In: Proceedingsofthe36th taskagents. arXivpreprintarXiv:2302.01560,2023
AnnualACMSymposiumonUserInterfaceSoftware 34. Lin J, Zhao H, Zhang A, Wu Y, Ping H, Chen Q.
andTechnology. 2023,1–22 Agentsims:Anopen-sourcesandboxforlargelanguage
21. Wang L, Zhang J, Chen X, Lin Y, Song R, Zhao model evaluation. arXiv preprint arXiv:2308.04026,
W X, Wen J R. Recagent: A novel simulation 2023
paradigm for recommender systems. arXiv preprint 35. LiangX,WangB,HuangH,WuS,WuP,LuL,MaZ,
arXiv:2306.02552,2023 Li Z. Unleashing infinite-length input capacity for
22. ZhangH,DuW,ShanJ,ZhouQ,DuY,Tenenbaum large-scalelanguagemodelswithself-controlledmem-
J B, Shu T, Gan C. Building cooperative embodied orysystem. arXivpreprintarXiv:2304.13343,2023
agentsmodularlywithlargelanguagemodels. arXiv 36. NgY,MiyashitaD,HoshiY,MoriokaY,ToriiO,Ko-
preprintarXiv:2307.02485,2023 dama T, Deguchi J. Simplyretrieve: A private and
23. HongS,ZhengX,ChenJ,ChengY,WangJ,ZhangC, lightweightretrieval-centricgenerativeaitool. arXiv
WangZ,YauSKS,LinZ,ZhouL,others. Metagpt: preprintarXiv:2308.03983,2023
Metaprogrammingformulti-agentcollaborativeframe- 37. HuangZ,GutierrezS,KamanaH,MacNeilS.Memory
work. arXivpreprintarXiv:2308.00352,2023 sandbox: Transparentandinteractivememorymanage-
24. Dong Y, Jiang X, Jin Z, Li G. Self-collaboration mentforconversationalagents. In: AdjunctProceed-