pith. sign in

arxiv: 2502.08691 · v2 · submitted 2025-02-12 · 💻 cs.SI · cs.AI

AgentSociety: Large-Scale Simulation of LLM-Driven Generative Agents Advances Understanding of Human Behaviors and Society

Jinghua Piao , Yuwei Yan , Jun Zhang , Nian Li , Junbo Yan , Xiaochong Lan , Zhihong Lu , Zhiheng Zheng
show 8 more authors
Jing Yi Wang Di Zhou Chen Gao Fengli Xu Fang Zhang Ke Rong Jun Su Yong Li
This is my paper

Pith reviewed 2026-05-23 04:03 UTC · model grok-4.3

classification 💻 cs.SI cs.AI
keywords large-scale social simulationLLM-driven generative agentscomputational social sciencepolarizationuniversal basic incomeexternal shocksurban sustainability
0
0 comments X

The pith

Large-scale simulator of LLM agents reproduces real-world results on polarization, UBI and shocks

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper introduces AgentSociety as a simulator built from LLM-driven agents placed in a realistic societal environment and run by a large-scale engine. It generates full social lives for more than 10,000 agents that produce five million interactions. The authors then apply the simulator to five social questions: polarization, spread of inflammatory messages, effects of universal basic income, impact of external shocks such as hurricanes, and urban sustainability. In each case the simulated outcomes line up with results from actual human experiments, which the authors take as evidence that the platform can support surveys, interviews and interventions for studying social mechanisms.

Core claim

AgentSociety integrates LLM-driven generative agents, a realistic societal environment, and a powerful large-scale simulation engine to generate social lives for over 10k agents through five million interactions; when applied to polarization, inflammatory message spread, universal basic income policies, hurricane shocks, and urban sustainability, the simulator produces outcomes that align with real-world experimental results.

What carries the argument

LLM-driven generative agents operating inside a realistic societal environment and driven by a large-scale simulation engine that records millions of agent-agent and agent-environment interactions.

If this is right

  • Researchers can apply standard social-science methods such as surveys, interviews, and interventions inside the simulation at scale.
  • The platform can be used to trace patterns, causes, and mechanisms behind the five tested social issues.
  • Social scientists and policymakers gain a replicable computational testbed for studying complex dynamics without the logistical costs of physical experiments.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Policymakers could run controlled tests of interventions such as universal basic income inside the simulator before any real-world deployment.
  • The same agent-environment setup could be extended to additional social phenomena beyond the five cases examined.
  • If the alignment with human data continues, the approach may lower the cost and increase the speed of exploratory social research.

Load-bearing premise

The behaviors and interactions of the LLM-driven agents in the simulated environment sufficiently mirror those of real humans to support conclusions about social dynamics.

What would settle it

A direct side-by-side comparison in which the simulator's measured effect of universal basic income on agent behavior or the measured impact of a simulated hurricane diverges substantially from the corresponding real-world experimental data.

Figures

Figures reproduced from arXiv: 2502.08691 by Chen Gao, Di Zhou, Fang Zhang, Fengli Xu, Jinghua Piao, Jing Yi Wang, Junbo Yan, Jun Su, Jun Zhang, Ke Rong, Nian Li, Xiaochong Lan, Yong Li, Yuwei Yan, Zhiheng Zheng, Zhihong Lu.

Figure 1
Figure 1. Figure 1: Evaluation framework for LLM-driven social simulators. [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Overview of the proposed social simulator AgentSociety. AgentSociety consists of three key [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Overview of LLM-driven social generative agents. [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Modeling framework of emotion, needs, and cognition. [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Modeling of mobility behavior. Mobility serves as a foundational behavioral module, operating as an integrative force within the social agent’s action network. This enables multidimensional coordination across social, economic, and environmental domains. The act of moving to a park, for instance, inherently carries the potential for social synergy. Spontaneous encounters with acquaintances may emerge, cata… view at source ↗
Figure 6
Figure 6. Figure 6: Modeling of social behavior. 10 [PITH_FULL_IMAGE:figures/full_fig_p010_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Modeling of economic behavior. In terms of behavior modeling, we simulate the employment and consumption behavior of social agents through the strength of their work and consumption propensity, and apply these behaviors in a macroeconomic simulation environment [64]. Work propensity determines the agent’s working hours and corresponding monthly income, while consumption propensity determines their monthly … view at source ↗
Figure 8
Figure 8. Figure 8: Workflow of social agents based on stream memory. [PITH_FULL_IMAGE:figures/full_fig_p013_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Overview of the societal environment. According to the introduction above, in the design and simulation of social agents, mobility behaviors, social behaviors, and economic behaviors like employment and consumption are essential external capabilities. In the real world, the manifestation of these behaviors is grounded in corresponding objective entities, not merely in human subjective cognition. For exampl… view at source ↗
Figure 10
Figure 10. Figure 10: System architecture of the large-scale social simulation engine. [PITH_FULL_IMAGE:figures/full_fig_p018_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Asynchronous multi-process parallel execution using Ray and asyncio. [PITH_FULL_IMAGE:figures/full_fig_p020_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Overview of MQTT-powered agent messaging system. [PITH_FULL_IMAGE:figures/full_fig_p020_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Distribution analysis for 10k agents. of LLM API performance and the trade-offs involved in private deployment options. To improve stability and scalability, further research should focus on optimizing the LLM infrastructure or exploring alternative solutions for large-scale intelligent agent simulations. 7 Exemplary Social Experiments 7.1 One Day Life This section presents a self-directed day in the life… view at source ↗
Figure 14
Figure 14. Figure 14: Experiment configuration overview [PITH_FULL_IMAGE:figures/full_fig_p028_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: Large-scale social simulation. 28 [PITH_FULL_IMAGE:figures/full_fig_p028_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: Opinion changes on the political issue of Gun Control across three experimental setups. [PITH_FULL_IMAGE:figures/full_fig_p030_16.png] view at source ↗
Figure 17
Figure 17. Figure 17: Simulation results of the spread of inflammatory messages. [PITH_FULL_IMAGE:figures/full_fig_p031_17.png] view at source ↗
Figure 18
Figure 18. Figure 18: Agent opinions on the chained woman incident. [PITH_FULL_IMAGE:figures/full_fig_p032_18.png] view at source ↗
Figure 19
Figure 19. Figure 19: Simulation results of the economic system. [PITH_FULL_IMAGE:figures/full_fig_p033_19.png] view at source ↗
Figure 20
Figure 20. Figure 20: The comparison of economic and social metrics. [PITH_FULL_IMAGE:figures/full_fig_p033_20.png] view at source ↗
Figure 21
Figure 21. Figure 21: Agent opinions on UBI policy. 7.5 External Shocks of Hurricane The impact of external disasters on human mobility is a critical area of study due to their profound effects on societal structures and individual behaviors. Understanding how such events influence human movement patterns is essential for enhancing emergency response strategies and mitigating potential risks. Hurricanes, as severe natural disa… view at source ↗
Figure 22
Figure 22. Figure 22: Activity level spatial distributions. The line graph presented above ( [PITH_FULL_IMAGE:figures/full_fig_p034_22.png] view at source ↗
Figure 23
Figure 23. Figure 23: Normalized daily trips. 8 Related Works The literature related to the work mainly consists of two kinds of work: large language model-driven agents and social simulation. 8.1 LLM-driven Agents Large Language Models (LLMs) exhibit astonishing language capabilities [99]. Since the language ability is one of the most fundamental abilities of human intelligence, LLMs demonstrate excellent performance in numer… view at source ↗
read the original abstract

Understanding human behavior and society is a central focus in social sciences, with the rise of generative social science marking a significant paradigmatic shift. By leveraging bottom-up simulations, it replaces costly and logistically challenging traditional experiments with scalable, replicable, and systematic computational approaches for studying complex social dynamics. Recent advances in large language models (LLMs) have further transformed this research paradigm, enabling the creation of human-like generative social agents and realistic simulacra of society. In this paper, we propose AgentSociety, a large-scale social simulator that integrates LLM-driven agents, a realistic societal environment, and a powerful large-scale simulation engine. Based on the proposed simulator, we generate social lives for over 10k agents, simulating their 5 million interactions both among agents and between agents and their environment. Furthermore, we explore the potential of AgentSociety as a testbed for computational social experiments, focusing on five key social issues: polarization, the spread of inflammatory messages, the effects of universal basic income policies, the impact of external shocks such as hurricanes, and urban sustainability. These five issues serve as valuable cases for assessing AgentSociety's support for typical research methods -- such as surveys, interviews, and interventions -- as well as for investigating the patterns, causes, and underlying mechanisms of social issues. The alignment between AgentSociety's outcomes and real-world experimental results not only demonstrates its ability to capture human behaviors and their underlying mechanisms, but also underscores its potential as an important platform for social scientists and policymakers.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 1 minor

Summary. The paper proposes AgentSociety, a large-scale social simulator integrating LLM-driven generative agents, a realistic societal environment, and a simulation engine. It generates social lives for over 10k agents with 5 million interactions and evaluates the platform on five social issues (polarization, spread of inflammatory messages, universal basic income policies, hurricane impacts, and urban sustainability) as test cases for surveys, interviews, and interventions. The central claim is that simulated outcomes align with real-world experimental results, demonstrating capture of human behaviors and underlying mechanisms and establishing the simulator as a platform for social scientists and policymakers.

Significance. If the alignment claims hold under rigorous quantitative validation with controls, the scale (10k agents, 5M interactions) and integration of LLM agents with an explicit environment could provide a valuable, replicable testbed for generative social science, enabling systematic study of complex dynamics that are costly to examine in the field. The work's potential rests on whether multi-agent interactions produce explanatory power beyond LLM priors.

major comments (3)
  1. [Abstract] Abstract: the claim that alignment with real-world experimental results 'demonstrates its ability to capture human behaviors and their underlying mechanisms' is unsupported by any quantitative metrics (e.g., KL divergence, effect sizes, prediction error), statistical tests, or experimental design details. This directly undermines the central validity claim for the five case studies.
  2. [Abstract] Abstract (five issues paragraph): no ablation or control is described that isolates the contribution of multi-agent interactions and the societal environment from the LLM's training corpus priors; without this, observed alignments could be consistent with surface reproduction rather than emergent simulation dynamics.
  3. [Abstract] Abstract: the text does not state whether simulation parameters were fitted to the target real-world data or run as parameter-free/out-of-sample predictions, which is required to evaluate whether the reported alignments reflect mechanism capture.
minor comments (1)
  1. [Abstract] Abstract: the phrase 'generate social lives for over 10k agents' would benefit from a parenthetical note on how agent initialization and environment realism are operationalized.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for these focused comments on the abstract and the strength of our central claims. We respond to each point below and indicate where revisions to the manuscript will be made.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the claim that alignment with real-world experimental results 'demonstrates its ability to capture human behaviors and their underlying mechanisms' is unsupported by any quantitative metrics (e.g., KL divergence, effect sizes, prediction error), statistical tests, or experimental design details. This directly undermines the central validity claim for the five case studies.

    Authors: We agree that the abstract phrasing is insufficiently precise on this point. The full manuscript contains quantitative comparisons (including effect-size alignments and statistical tests against real-world benchmarks) in the case-study sections, but these details are not referenced in the abstract. We will revise the abstract to cite the specific metrics and evaluation procedures used, thereby supporting the validity claim more rigorously. revision: yes

  2. Referee: [Abstract] Abstract (five issues paragraph): no ablation or control is described that isolates the contribution of multi-agent interactions and the societal environment from the LLM's training corpus priors; without this, observed alignments could be consistent with surface reproduction rather than emergent simulation dynamics.

    Authors: This observation is correct for the current abstract. While the manuscript architecture separates the LLM priors from the explicit environment and interaction rules, no explicit ablation isolating these components is described at the abstract level. We will add a concise statement of the control design (comparing full multi-agent runs against single-agent and environment-ablated baselines) to the revised abstract and expand the corresponding analysis in the methods section. revision: yes

  3. Referee: [Abstract] Abstract: the text does not state whether simulation parameters were fitted to the target real-world data or run as parameter-free/out-of-sample predictions, which is required to evaluate whether the reported alignments reflect mechanism capture.

    Authors: We accept the need for explicit clarification. All reported simulations were executed as parameter-free, out-of-sample predictions; no parameters were tuned to the target real-world datasets. We will revise the abstract to state this design choice directly, thereby strengthening the interpretation that observed alignments arise from captured mechanisms rather than data fitting. revision: yes

Circularity Check

0 steps flagged

No circularity: validation rests on external real-world benchmarks with no fitted predictions or self-referential definitions

full rationale

The paper describes a simulation platform (AgentSociety) and reports alignment of its outputs with five external real-world experimental results on polarization, inflammatory messages, UBI, hurricanes, and urban sustainability. No equations, parameter-fitting procedures, or derivation steps are presented in the abstract or described claims that would reduce any 'prediction' to the simulator's own inputs by construction. No self-citations are invoked as load-bearing uniqueness theorems, and no ansatz or renaming of known results is used to justify core claims. The central assertion of mechanism capture is therefore independent of the paper's own fitted values and stands as an empirical claim against outside benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 1 invented entities

Based on abstract; no specific free parameters mentioned. The core assumption is about the fidelity of LLM agents to human behavior.

axioms (1)
  • domain assumption LLM-driven agents can generate realistic human-like social behaviors and interactions
    This is required for the claim that simulation outcomes align with and capture real human behaviors.
invented entities (1)
  • AgentSociety simulator no independent evidence
    purpose: Large-scale simulation of societal dynamics using LLM agents
    Newly introduced platform in the paper.

pith-pipeline@v0.9.0 · 5856 in / 1170 out tokens · 64965 ms · 2026-05-23T04:03:24.291500+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 26 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. AgentSocialBench: Evaluating Privacy Risks in Human-Centered Agentic Social Networks

    cs.AI 2026-04 unverdicted novelty 8.0

    AgentSocialBench demonstrates that privacy preservation is fundamentally harder in human-centered agentic social networks than in single-agent cases due to cross-domain coordination pressures and an abstraction parado...

  2. Can LLMs Think Like Consumers? Benchmarking Crowd-Level Reaction Reconstruction with ConsumerSimBench

    cs.CL 2026-05 unverdicted novelty 7.0

    ConsumerSimBench evaluates 13 LLMs on reconstructing crowd reactions from 1,553 Chinese social-media topics using 23,122 auditable yes-no criteria, finding maximum coverage of 47.8% by Gemini-3.1-Pro.

  3. ScioMind: Cognitively Grounded Multi-Agent Social Simulation with Anchoring-Based Belief Dynamics and Dynamic Profiles

    cs.AI 2026-05 unverdicted novelty 7.0

    ScioMind combines anchoring-based belief updates, hierarchical memory, and dynamic profiles in LLM multi-agent systems to produce more stable, diverse, and psychologically aligned opinion trajectories than prior fixed...

  4. Agentic World Modeling: Foundations, Capabilities, Laws, and Beyond

    cs.AI 2026-04 unverdicted novelty 7.0

    Proposes a levels x laws taxonomy for world models in AI agents, defining L1-L3 capabilities across physical, digital, social, and scientific regimes while reviewing over 400 works to outline a roadmap for advanced ag...

  5. AI-Gram: When Visual Agents Interact in a Social Network

    cs.AI 2026-04 unverdicted novelty 7.0

    Autonomous visual AI agents spontaneously form image reply chains, maintain stable individual styles, and produce richer style-diverse conversations than single agents can achieve alone.

  6. Memory-Augmented LLM-based Multi-Agent System for Automated Feature Generation on Tabular Data

    cs.AI 2026-04 unverdicted novelty 7.0

    MALMAS is a memory-augmented multi-agent LLM system that generates diverse, high-quality features for tabular data via agent decomposition, routing, and iterative memory-guided refinement.

  7. WhatIf: Interactive Exploration of LLM-Powered Social Simulations for Policy Reasoning

    cs.HC 2026-04 unverdicted novelty 7.0

    WhatIf provides an interactive platform for real-time exploration of LLM-driven social simulations, enabling policymakers to iteratively test plans, reflect on assumptions, and uncover vulnerabilities in emergency pre...

  8. IntervenSim: Intervention-Aware Social Network Simulation for Opinion Dynamics

    cs.SI 2026-04 unverdicted novelty 7.0

    IntervenSim is an intervention-aware social network simulation that couples source interventions with crowd interactions in a feedback loop, improving MAPE by 41.6% and DTW by 66.9% over prior static frameworks on rea...

  9. Market-Bench: Benchmarking Large Language Models on Economic and Trade Competition

    cs.AI 2026-04 conditional novelty 7.0

    Market-Bench is a new multi-agent benchmark showing that LLMs display large performance gaps in economic tasks, with only a few consistently growing capital while most break even despite similar ad quality.

  10. GRAPHIA: Harnessing Social Graph Data to Enhance LLM-Based Social Simulation

    cs.SI 2025-10 unverdicted novelty 7.0

    Graphia is an LLM post-training framework that uses real social graphs and GNN rewards to improve micro-level interaction prediction and macro-level network property replication in dynamic social simulations.

  11. The PIMMUR Principles: Ensuring Validity in Collective Behavior of LLM Societies

    cs.CL 2025-09 conditional novelty 7.0

    A systematic audit of LLM-based AI societies finds that 89.7% of 39 studies violate at least one of six PIMMUR validity principles, with reproductions showing that many claimed collective behaviors disappear when cont...

  12. Stop Drawing Scientific Claims from LLM Social Simulations Without Robustness Audits

    physics.soc-ph 2026-05 accept novelty 6.0

    Minor perturbations in persona format, instruction framing, and network structure shift cooperation by up to 76 percentage points and polarization metrics consistently, showing that LLM social simulations require per-...

  13. A Survey on LLM-based Conversational User Simulation

    cs.CL 2026-04 unverdicted novelty 6.0

    A survey that introduces a taxonomy for LLM-based conversational user simulation, analyzes core techniques and evaluation methods, and identifies open challenges in the field.

  14. Superminds Test: Actively Evaluating Collective Intelligence of Agent Society via Probing Agents

    cs.AI 2026-04 unverdicted novelty 6.0

    Large-scale experiments on two million agents reveal that collective intelligence does not emerge from scale alone due to sparse and shallow interactions.

  15. Topology-Aware LLM-Driven Social Simulation: A Unified Framework for Efficient and Realistic Agent Dynamics

    cs.SI 2026-04 unverdicted novelty 6.0

    TopoSim integrates network topology into LLM agent simulations via backbone units and heterogeneous influence to cut token use 50-90% while improving fidelity to real-world structures.

  16. TrafficClaw: Generalizable Urban Traffic Control via Unified Physical Environment Modeling

    cs.AI 2026-04 unverdicted novelty 6.0

    TrafficClaw creates a single runtime environment for heterogeneous urban traffic subsystems and deploys an LLM agent with spatiotemporal reasoning to deliver robust control that generalizes across unseen scenarios.

  17. EvoSpark: Endogenous Interactive Agent Societies for Unified Long-Horizon Narrative Evolution

    cs.CL 2026-04 unverdicted novelty 6.0

    EvoSpark introduces Stratified Narrative Memory, Generative Mise-en-Scène, and a Unified Narrative Operation Engine to sustain coherent long-horizon narratives in LLM-based multi-agent societies.

  18. Evaluating Cooperation in LLM Social Groups through Elected Leadership

    cs.CL 2026-04 unverdicted novelty 6.0

    Elected leadership in LLM multi-agent simulations of common-pool resource governance raises social welfare scores by 55.4% and survival time by 128.6%.

  19. When simulations look right but causal effects go wrong: Large language models as behavioral simulators

    cs.CY 2026-04 unverdicted novelty 6.0

    LLMs reproduce observed attitudinal patterns in climate interventions reasonably well but diverge on causal effect estimates, with descriptive fit failing to predict causal accuracy across interventions and outcomes.

  20. A Survey on Generative Recommendation: Data, Model, and Tasks

    cs.IR 2025-10 accept novelty 6.0

    This survey organizes generative recommendation into data, model, and task dimensions, identifying five advantages including world knowledge integration and creative generation while noting challenges in benchmarks an...

  21. TokenCake: A KV-Cache-centric Serving Framework for LLM-based Multi-Agent Applications

    cs.DC 2025-10 unverdicted novelty 6.0

    TokenCake introduces agent-aware temporal and spatial schedulers for KV cache management in LLM multi-agent serving, claiming over 47% lower end-to-end latency and up to 16.9% better GPU memory utilization than vLLM o...

  22. Do LLM Agents Mirror Socio-Cognitive Effects in Power-Asymmetric Conversations?

    cs.CL 2026-05 unverdicted novelty 5.0

    LLM agents in power-asymmetric role-play exhibit socio-cognitive effects including linguistic coordination, pronoun usage patterns, persuasion success, and compliance with unsafe requests.

  23. Do LLM Agents Mirror Socio-Cognitive Effects in Power-Asymmetric Conversations?

    cs.CL 2026-05 unverdicted novelty 5.0

    LLMs assigned high or low status personas in multi-turn dialogues exhibit socio-cognitive effects including language coordination, pronoun patterns, persuasion success, and compliance with unsafe requests.

  24. AgentDynEx: Nudging the Mechanics and Dynamics of Multi-Agent Simulations

    cs.MA 2025-04 unverdicted novelty 5.0

    AgentDynEx introduces nudging and a Configuration Matrix to help set up and maintain balanced mechanics and dynamics in multi-agent LLM simulations.

  25. Agentic Microphysics: A Manifesto for Generative AI Safety

    cs.CY 2026-04 unverdicted novelty 4.0

    The authors introduce agentic microphysics and generative safety to link local agent interactions to population-level risks in agentic AI through a causally explicit framework.

  26. From Script to Stage: Automating Experimental Design for Social Simulations with LLMs

    cs.HC 2025-10 unverdicted novelty 4.0

    FSTS automates multi-agent social experiment design via LLM script generation across three phases, with tests indicating reproduction of real-world outcomes.

Reference graph

Works this paper leans on

118 extracted references · 118 canonical work pages · cited by 25 Pith papers · 10 internal anchors

  1. [1]

    Perils and oppor- tunities in using large language models in psychological research

    Suhaib Abdurahman, Mohammad Atari, Farzan Karimi-Malekabadi, Mona J Xue, Jackson Trager, Peter S Park, Preni Golazizian, Ali Omrani, and Morteza Dehghani. Perils and oppor- tunities in using large language models in psychological research. PNAS nexus, 3(7):pgae245, 2024

    work page 2024

  2. [2]

    Large language models show human-like content biases in transmission chain experiments

    Alberto Acerbi and Joseph M Stubbersfield. Large language models show human-like content biases in transmission chain experiments. Proceedings of the National Academy of Sciences, 120(44):e2313790120, 2023

    work page 2023

  3. [3]

    humanistic

    Alma Acevedo. A personalistic appraisal of maslow’s needs theory of motivation: From “humanistic” psychology to integral humanism. Journal of business ethics , 148:741–763, 2018

    work page 2018

  4. [4]

    Using large language models to simulate multiple humans and replicate human subject studies

    Gati V Aher, Rosa I Arriaga, and Adam Tauman Kalai. Using large language models to simulate multiple humans and replicate human subject studies. In International Conference on Machine Learning, pages 337–371. PMLR, 2023

    work page 2023

  5. [5]

    The theory of planned behavior

    Icek Ajzen. The theory of planned behavior. Organizational Behavior and Human Decision Processes, 1991. 39

    work page 1991

  6. [6]

    Project sid: Many-agent simulations toward ai civilization

    Altera AL, Andrew Ahn, Nic Becker, Stephanie Carroll, Nico Christie, Manuel Cortes, Arda Demirci, Melissa Du, Frankie Li, Shuying Luo, et al. Project sid: Many-agent simulations toward ai civilization. arXiv preprint arXiv:2411.00114, 2024

    work page arXiv 2024

  7. [7]

    Chatgpt and the rise of semi-humans

    Abdulrahman Essa Al Lily, Abdelrahim Fathy Ismail, Fathi M Abunaser, Firass Al-Lami, and Ali Khalifa Atwa Abdullatif. Chatgpt and the rise of semi-humans. Humanities and Social Sciences Communications, 10(1):1–12, 2023

    work page 2023

  8. [8]

    Challenges, tasks, and opportunities in modeling agent-based complex systems

    Li An, V olker Grimm, Abigail Sullivan, BL Turner Ii, Nicolas Malleson, Alison Heppenstall, Christian Vincenot, Derek Robinson, Xinyue Ye, Jianguo Liu, et al. Challenges, tasks, and opportunities in modeling agent-based complex systems. Ecological Modelling, 457:109685, 2021

    work page 2021

  9. [9]

    Out of one, many: Using language models to simulate human samples

    Lisa P Argyle, Ethan C Busby, Nancy Fulda, Joshua R Gubler, Christopher Rytting, and David Wingate. Out of one, many: Using language models to simulate human samples. Political Analysis, 31(3):337–351, 2023

    work page 2023

  10. [10]

    Out-of-equilibrium economics and agent-based modeling

    W Brian Arthur. Out-of-equilibrium economics and agent-based modeling. Handbook of computational economics, 2:1551–1564, 2006

    work page 2006

  11. [11]

    Predicting results of social science experiments using large language models

    Ashwini Ashokkumar, Luke Hewitt, Isaias Ghezae, and Robb Willer. Predicting results of social science experiments using large language models. Work. Pap., New York Univ., New York, NY, 2024

    work page 2024

  12. [12]

    Agent-based modeling in economics and finance: Past, present, and future

    Robert L Axtell and J Doyne Farmer. Agent-based modeling in economics and finance: Past, present, and future. Journal of Economic Literature, pages 1–101, 2022

    work page 2022

  13. [13]

    Human agency in social cognitive theory

    Albert Bandura. Human agency in social cognitive theory. American psychologist, 44(9):1175, 1989

    work page 1989

  14. [14]

    The impact of unconditional cash transfers on consumption and household balance sheets: Experimental evidence from two us states

    Alexander W Bartik, Elizabeth Rhodes, David E Broockman, Patrick K Krause, Sarah Miller, and Eva Vivalt. The impact of unconditional cash transfers on consumption and household balance sheets: Experimental evidence from two us states. Technical report, National Bureau of Economic Research, 2024

    work page 2024

  15. [15]

    Emergence of polarized ideological opinions in multidimensional topic spaces

    Fabian Baumann, Philipp Lorenz-Spreen, Igor M Sokolov, and Michele Starnini. Emergence of polarized ideological opinions in multidimensional topic spaces. Physical Review X , 11(1):011012, 2021

    work page 2021

  16. [16]

    Emotivational psychology: How distinct emotions facilitate fundamental motives

    Alec T Beall and Jessica L Tracy. Emotivational psychology: How distinct emotions facilitate fundamental motives. Social and Personality Psychology Compass, 11(2):e12303, 2017

    work page 2017

  17. [17]

    Sumo–simulation of urban mobility: an overview

    Michael Behrisch, Laura Bieker, Jakob Erdmann, and Daniel Krajzewicz. Sumo–simulation of urban mobility: an overview. In Proceedings of SIMUL 2011, The Third International Conference on Advances in System Simulation. ThinkMind, 2011

    work page 2011

  18. [18]

    Adaptive agents, intelligence, and emergent human organization: Capturing complexity through agent-based modeling

    Brian JL Berry, L Douglas Kiel, and Euel Elliott. Adaptive agents, intelligence, and emergent human organization: Capturing complexity through agent-based modeling. Proceedings of the National Academy of Sciences, 99(suppl_3):7187–7188, 2002

    work page 2002

  19. [19]

    Autonomous chemical research with large language models

    Daniil A Boiko, Robert MacKnight, Ben Kline, and Gabe Gomes. Autonomous chemical research with large language models. Nature, 624(7992):570–578, 2023

    work page 2023

  20. [20]

    Emotion modeling in social simulation: a survey

    Mathieu Bourgais, Patrick Taillandier, Laurent Vercouter, and Carole Adam. Emotion modeling in social simulation: a survey. Journal of Artificial Societies and Social Simulation, 2018

    work page 2018

  21. [21]

    Emotion shapes the diffusion of moralized content in social networks

    William J Brady, Julian A Wills, John T Jost, Joshua A Tucker, and Jay J Van Bavel. Emotion shapes the diffusion of moralized content in social networks. Proceedings of the National Academy of Sciences, 114(28):7313–7318, 2017

    work page 2017

  22. [22]

    Statistical modeling: The two cultures (with comments and a rejoinder by the author)

    Leo Breiman. Statistical modeling: The two cultures (with comments and a rejoinder by the author). Statistical science, 16(3):199–231, 2001

    work page 2001

  23. [23]

    Agent-based economic models and econometrics

    Shu-Heng Chen, Chia-Ling Chang, and Ye-Rong Du. Agent-based economic models and econometrics. The Knowledge Engineering Review, 27(2):187–219, 2012

    work page 2012

  24. [24]

    Sociodojo: Building lifelong analytical agents with real-world text and time series

    Junyan Cheng and Peter Chin. Sociodojo: Building lifelong analytical agents with real-world text and time series. In The Twelfth International Conference on Learning Representations, 2024. 40

    work page 2024

  25. [25]

    Nominal rigidities and the dynamic effects of a shock to monetary policy

    Lawrence J Christiano, Martin Eichenbaum, and Charles L Evans. Nominal rigidities and the dynamic effects of a shock to monetary policy. Journal of political Economy, 113(1):1–45, 2005

    work page 2005

  26. [26]

    The game of life

    John Conway et al. The game of life. Scientific American, 223(4):4, 1970

    work page 1970

  27. [27]

    Agent-based models

    Scott De Marchi and Scott E Page. Agent-based models. Annual Review of political science, 17(1):1–20, 2014

    work page 2014

  28. [28]

    Using large language models in psychology

    Dorottya Demszky, Diyi Yang, David S Yeager, Christopher J Bryan, Margarett Clapper, Susannah Chandhok, Johannes C Eichstaedt, Cameron Hecht, Jeremy Jamieson, Meghann Johnson, et al. Using large language models in psychology. Nature Reviews Psychology, 2(11):688–701, 2023

    work page 2023

  29. [29]

    Growing Artificial Societies: Social Science from the Bottom Up

    Joshua M Epstein. Growing Artificial Societies: Social Science from the Bottom Up . The Brookings Institution Press, 1996

    work page 1996

  30. [30]

    Agent-based computational models and generative social science

    Joshua M Epstein. Agent-based computational models and generative social science. Com- plexity, 4(5):41–60, 1999

    work page 1999

  31. [31]

    Generative social science: Studies in agent-based computational modeling

    Joshua M Epstein. Generative social science: Studies in agent-based computational modeling. Princeton University Press, 2012

    work page 2012

  32. [32]

    Cognitive psychology: A student’s handbook

    Michael W Eysenck and Mark T Keane. Cognitive psychology: A student’s handbook . Psychology press, 2020

    work page 2020

  33. [33]

    Agentmove: Predicting human mobility anywhere using large language model based agentic framework

    Jie Feng, Yuwei Du, Jie Zhao, and Yong Li. Agentmove: Predicting human mobility anywhere using large language model based agentic framework. arXiv preprint arXiv:2408.13986, 2024

    work page arXiv 2024

  34. [34]

    Linking agent- based models and stochastic models of financial markets.Proceedings of the National Academy of Sciences, 109(22):8388–8393, 2012

    Ling Feng, Baowen Li, Boris Podobnik, Tobias Preis, and H Eugene Stanley. Linking agent- based models and stochastic models of financial markets.Proceedings of the National Academy of Sciences, 109(22):8388–8393, 2012

    work page 2012

  35. [35]

    Intelligent driving intelligence test for autonomous vehicles with naturalistic and adversarial environment.Nature communications, 12(1):748, 2021

    Shuo Feng, Xintao Yan, Haowei Sun, Yiheng Feng, and Henry X Liu. Intelligent driving intelligence test for autonomous vehicles with naturalistic and adversarial environment.Nature communications, 12(1):748, 2021

    work page 2021

  36. [36]

    Empirical evidence using microsimulation models in the social sciences

    Francesco Figari, Emanuela Lezzi, et al. Empirical evidence using microsimulation models in the social sciences. New Horizons in Modeling and Simulation for Social Epidemiology and Public Health, pages 107–148, 2021

    work page 2021

  37. [37]

    Reconstructing economics: Agent based models and complexity

    Mauro Gallegati and Alan Kirman. Reconstructing economics: Agent based models and complexity. Complexity Economics, 1(1):5–31, 2012

    work page 2012

  38. [38]

    Large language models empowered agent-based modeling and simulation: A survey and perspectives

    Chen Gao, Xiaochong Lan, Nian Li, Yuan Yuan, Jingtao Ding, Zhilun Zhou, Fengli Xu, and Yong Li. Large language models empowered agent-based modeling and simulation: A survey and perspectives. Humanities and Social Sciences Communications, 11(1):1–24, 2024

    work page 2024

  39. [39]

    S$^3$: Social-network Simulation System with Large Language Model-Empowered Agents

    Chen Gao, Xiaochong Lan, Zhihong Lu, Jinzhu Mao, Jinghua Piao, Huandong Wang, Depeng Jin, and Yong Li. S3: Social-network simulation system with large language model-empowered agents. arXiv preprint arXiv:2307.14984, 2023

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  40. [40]

    AgentScope: A Flexible yet Robust Multi-Agent Platform,

    Dawei Gao, Zitao Li, Xuchen Pan, Weirui Kuang, Zhijian Ma, Bingchen Qian, Fei Wei, Wenhao Zhang, Yuexiang Xie, Daoyuan Chen, et al. Agentscope: A flexible yet robust multi-agent platform. arXiv preprint arXiv:2402.14034, 2024

    work page arXiv 2024

  41. [41]

    How to build and use agent-based models in social science

    Nigel Gilbert and Pietro Terna. How to build and use agent-based models in social science. Mind & Society, 1:57–72, 2000

    work page 2000

  42. [42]

    Computational models of collective behavior

    Robert L Goldstone and Marco A Janssen. Computational models of collective behavior. Trends in cognitive sciences, 9(9):424–430, 2005

    work page 2005

  43. [43]

    Empowering working memory for large language model agents

    Jing Guo, Nan Li, Jianchuan Qi, Hang Yang, Ruiqiao Li, Yuzhen Feng, Si Zhang, and Ming Xu. Empowering working memory for large language model agents. arXiv preprint arXiv:2312.17259, 2023

    work page arXiv 2023

  44. [44]

    Causal mechanisms in the social sciences

    Peter Hedström and Petri Ylikoski. Causal mechanisms in the social sciences. Annual review of sociology, 36(1):49–67, 2010

    work page 2010

  45. [45]

    Social force model for pedestrian dynamics

    Dirk Helbing and Peter Molnar. Social force model for pedestrian dynamics. Physical review E, 51(5):4282, 1995. 41

    work page 1995

  46. [46]

    Integrating explanation and prediction in computational social science

    Jake M Hofman, Duncan J Watts, Susan Athey, Filiz Garip, Thomas L Griffiths, Jon Kleinberg, Helen Margetts, Sendhil Mullainathan, Matthew J Salganik, Simine Vazire, et al. Integrating explanation and prediction in computational social science. Nature, 595(7866):181–188, 2021

    work page 2021

  47. [47]

    MetaGPT: Meta Programming for A Multi-Agent Collaborative Framework

    Sirui Hong, Xiawu Zheng, Jonathan Chen, Yuheng Cheng, Jinlin Wang, Ceyao Zhang, Zili Wang, Steven Ka Shing Yau, Zijuan Lin, Liyang Zhou, et al. Metagpt: Meta programming for multi-agent collaborative framework. arXiv preprint arXiv:2308.00352, 2023

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  48. [48]

    Large language models as simulated economic agents: What can we learn from homo silicus? Technical report, National Bureau of Economic Research, 2023

    John J Horton. Large language models as simulated economic agents: What can we learn from homo silicus? Technical report, National Bureau of Economic Research, 2023

    work page 2023

  49. [49]

    A survey on hallucination in large language models: Principles, taxonomy, challenges, and open questions

    Lei Huang, Weijiang Yu, Weitao Ma, Weihong Zhong, Zhangyin Feng, Haotian Wang, Qiang- long Chen, Weihua Peng, Xiaocheng Feng, Bing Qin, et al. A survey on hallucination in large language models: Principles, taxonomy, challenges, and open questions. ACM Transactions on Information Systems, 2023

    work page 2023

  50. [50]

    Benchmarking large language models as ai research agents

    Qian Huang, Jian V ora, Percy Liang, and Jure Leskovec. Benchmarking large language models as ai research agents. In NeurIPS 2023 Foundation Models for Decision Making Workshop, 2023

    work page 2023

  51. [51]

    Social science meets llms: How reliable are large language models in social simulations? arXiv preprint arXiv:2410.23426, 2024

    Yue Huang, Zhengqing Yuan, Yujun Zhou, Kehan Guo, Xiangqi Wang, Haomin Zhuang, Weixiang Sun, Lichao Sun, Jindong Wang, Yanfang Ye, et al. Social science meets llms: How reliable are large language models in social simulations? arXiv preprint arXiv:2410.23426, 2024

    work page arXiv 2024

  52. [52]

    Casevo: A cognitive agents and social evolution simulator

    Zexun Jiang, Yafang Shi, Maoxu Li, Hongjiang Xiao, Yunxiao Qin, Qinglan Wei, Ye Wang, and Yuan Zhang. Casevo: A cognitive agents and social evolution simulator. arXiv preprint arXiv:2412.19498, 2024

    work page arXiv 2024

  53. [53]

    Big data, agents, and machine learning: towards a data-driven agent-based modeling approach

    Hamdi Kavak, Jose J Padilla, Christopher J Lynch, and Saikou Y Diallo. Big data, agents, and machine learning: towards a data-driven agent-based modeling approach. In Proceedings of the Annual Simulation Symposium, pages 1–12, 2018

    work page 2018

  54. [54]

    General lane-changing model mobil for car-following models

    Arne Kesting, Martin Treiber, and Dirk Helbing. General lane-changing model mobil for car-following models. Transportation Research Record, 1999(1):86–94, 2007

    work page 1999

  55. [55]

    Evaluating large language models in theory of mind tasks

    Michal Kosinski. Evaluating large language models in theory of mind tasks. Proceedings of the National Academy of Sciences, 121(45):e2405460121, 2024

    work page 2024

  56. [56]

    Efficient memory management for large language model serving with pagedattention

    Woosuk Kwon, Zhuohan Li, Siyuan Zhuang, Ying Sheng, Lianmin Zheng, Cody Hao Yu, Joseph Gonzalez, Hao Zhang, and Ion Stoica. Efficient memory management for large language model serving with pagedattention. In Proceedings of the 29th Symposium on Operating Systems Principles, pages 611–626, 2023

    work page 2023

  57. [57]

    What is a complex system? European Journal for Philosophy of Science, 3:33–67, 2013

    James Ladyman, James Lambert, and Karoline Wiesner. What is a complex system? European Journal for Philosophy of Science, 3:33–67, 2013

    work page 2013

  58. [58]

    Evolving ai collectives to enhance human diversity and enable self-regulation

    Shiyang Lai, Yujin Potter, Junsol Kim, Richard Zhuang, Dawn Song, and James Evans. Evolving ai collectives to enhance human diversity and enable self-regulation. arXiv preprint arXiv:2402.12590, 2024

    work page arXiv 2024

  59. [59]

    Language models, like humans, show content effects on reasoning tasks

    Andrew K Lampinen, Ishita Dasgupta, Stephanie CY Chan, Hannah R Sheahan, Antonia Creswell, Dharshan Kumaran, James L McClelland, and Felix Hill. Language models, like humans, show content effects on reasoning tasks. PNAS nexus, 3(7):pgae233, 2024

    work page 2024

  60. [60]

    Party competition: An agent-based model, volume 20

    Michael Laver and Ernest Sergenti. Party competition: An agent-based model, volume 20. Princeton University Press, 2011

    work page 2011

  61. [61]

    Computational social science

    David Lazer, Alex Pentland, Lada Adamic, Sinan Aral, Albert-László Barabási, Devon Brewer, Nicholas Christakis, Noshir Contractor, James Fowler, Myron Gutmann, et al. Computational social science. Science, 323(5915):721–723, 2009

    work page 2009

  62. [62]

    Computational social science: Obstacles and opportunities

    David MJ Lazer, Alex Pentland, Duncan J Watts, Sinan Aral, Susan Athey, Noshir Contractor, Deen Freelon, Sandra Gonzalez-Bailon, Gary King, Helen Margetts, et al. Computational social science: Obstacles and opportunities. Science, 369(6507):1060–1062, 2020

    work page 2020

  63. [63]

    Camel: Communicative agents for" mind" exploration of large language model society

    Guohao Li, Hasan Hammoud, Hani Itani, Dmitrii Khizbullin, and Bernard Ghanem. Camel: Communicative agents for" mind" exploration of large language model society. Advances in Neural Information Processing Systems, 36:51991–52008, 2023. 42

    work page 2023

  64. [64]

    Econagent: large language model-empowered agents for simulating macroeconomic activities

    Nian Li, Chen Gao, Mingyu Li, Yong Li, and Qingmin Liao. Econagent: large language model-empowered agents for simulating macroeconomic activities. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 15523–15536, 2024

    work page 2024

  65. [65]

    DeepSeek-V3 Technical Report

    Aixin Liu, Bei Feng, Bing Xue, Bingxuan Wang, Bochao Wu, Chengda Lu, Chenggang Zhao, Chengqi Deng, Chenyu Zhang, Chong Ruan, et al. Deepseek-v3 technical report. arXiv preprint arXiv:2412.19437, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  66. [66]

    M., Yang, D., and V osoughi, S

    Ruibo Liu, Ruixin Yang, Chenyan Jia, Ge Zhang, Denny Zhou, Andrew M Dai, Diyi Yang, and Soroush V osoughi. Training socially aligned language models on simulated social interactions. arXiv preprint arXiv:2305.16960, 2023

    work page arXiv 2023

  67. [67]

    Tutorial on agent-based modeling and simulation

    Charles M Macal and Michael J North. Tutorial on agent-based modeling and simulation. In Proceedings of the Winter Simulation Conference, 2005., pages 14–pp. IEEE, 2005

    work page 2005

  68. [68]

    A theory of human motivation

    AH Maslow. A theory of human motivation. Psychological Review, 2:21–28, 1943

    work page 1943

  69. [69]

    Maslow’s hierarchy of needs

    Saul McLeod. Maslow’s hierarchy of needs. Simply psychology, 1(1-18), 2007

    work page 2007

  70. [70]

    Mind, self, and society from the standpoint of a social behaviorist

    George Herbert Mead. Mind, self, and society from the standpoint of a social behaviorist. Chicago, 1934

    work page 1934

  71. [71]

    Ray: A distributed framework for emerging {AI} applications

    Philipp Moritz, Robert Nishihara, Stephanie Wang, Alexey Tumanov, Richard Liaw, Eric Liang, Melih Elibol, Zongheng Yang, William Paul, Michael I Jordan, et al. Ray: A distributed framework for emerging {AI} applications. In 13th USENIX symposium on operating systems design and implementation (OSDI 18), pages 561–577, 2018

    work page 2018

  72. [72]

    Unveiling the truth and facilitating change: To- wards agent-based large-scale social movement simulation

    Xinyi Mou, Zhongyu Wei, and Xuanjing Huang. Unveiling the truth and facilitating change: To- wards agent-based large-scale social movement simulation. arXiv preprint arXiv:2402.16333, 2024

    work page arXiv 2024

  73. [73]

    Agent-based modeling and network dynamics

    Akira Namatame and Shu-Heng Chen. Agent-based modeling and network dynamics. Oxford University Press, 2016

    work page 2016

  74. [74]

    Human-like problem-solving abilities in large language models using chatgpt

    Graziella Orrù, Andrea Piarulli, Ciro Conversano, and Angelo Gemignani. Human-like problem-solving abilities in large language models using chatgpt. Frontiers in artificial intelligence, 6:1199350, 2023

    work page 2023

  75. [75]

    Self-alignment of large language models via monopolylogue-based social scene simulation

    Xianghe Pang, Shuo Tang, Rui Ye, Yuxin Xiong, Bolun Zhang, Yanfeng Wang, and Si- heng Chen. Self-alignment of large language models via monopolylogue-based social scene simulation. In Forty-first International Conference on Machine Learning, 2024

    work page 2024

  76. [76]

    Self-alignment of large language models via multi-agent social simulation

    Xianghe Pang, Shuo Tang, Rui Ye, Yuxin Xiong, Bolun Zhang, Yanfeng Wang, and Siheng Chen. Self-alignment of large language models via multi-agent social simulation. In ICLR 2024 Workshop on Large Language Model (LLM) Agents, 2024

    work page 2024

  77. [77]

    Generative agents: Interactive simulacra of human behavior

    Joon Sung Park, Joseph O’Brien, Carrie Jun Cai, Meredith Ringel Morris, Percy Liang, and Michael S Bernstein. Generative agents: Interactive simulacra of human behavior. In Proceedings of the 36th annual acm symposium on user interface software and technology, pages 1–22, 2023

    work page 2023

  78. [78]

    LLM Agents Grounded in Self-Reports Enable General-Purpose Simulation of Individuals

    Joon Sung Park, Carolyn Q Zou, Aaron Shaw, Benjamin Mako Hill, Carrie Cai, Mered- ith Ringel Morris, Robb Willer, Percy Liang, and Michael S Bernstein. Generative agent simulations of 1,000 people. arXiv preprint arXiv:2411.10109, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  79. [79]

    Human–ai adaptive dynamics drives the emergence of information cocoons

    Jinghua Piao, Jiazhen Liu, Fang Zhang, Jun Su, and Yong Li. Human–ai adaptive dynamics drives the emergence of information cocoons. Nature Machine Intelligence, 5(11):1214–1224, 2023

    work page 2023

  80. [80]

    Emergence of human-like polarization among large language model agents

    Jinghua Piao, Zhihong Lu, Chen Gao, Fengli Xu, Fernando P Santos, Yong Li, and James Evans. Emergence of human-like polarization among large language model agents. arXiv preprint arXiv:2501.05171, 2025

    work page arXiv 2025

Showing first 80 references.