Pith. sign in

REVIEW 4 major objections 5 minor 3 cited by

Coached LLM agents develop stable interaction patterns and form emergent social ties, yielding networks whose density, clustering, and path length resemble real online communities.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Coached LLM agents rewarded for social behaviors form interaction networks whose density, clustering, and path lengths fall near ranges seen in real online social graphs.

T0 review reviewed 2026-08-04 challenge →

load-bearing objection A reasonable simulation framework, but the 'emergent learning' claim is undermined by the Plan Prompt's explicit behavioral rules; needs an ablation and proper stats. the 4 major comments →

arxiv 2510.19299 v2 pith:T3LP2B4E submitted 2025-10-22 cs.AI cs.MAcs.SI

Learning to Make Friends: Coaching LLM Agents toward Emergent Social Ties

Philipp J. Schneider , Lin Tian , Marian-Andrei Rizoiu This is my paper
classification cs.AI cs.MAcs.SI
keywords multi-agent LLM simulationemergent social tiesin-context learningbehavioral rewardssocial network analysisonline communitiescoaching signalagent-based modeling
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The reading

This paper attempts to show that a population of large language model agents, each nudged by a set of behavioral rewards and an optional coaching tip, can build a social network from scratch rather than following a pre-defined graph. Across repeated rounds of posting, commenting, direct messaging, voting, and reflecting, the agents develop stable directed ties, and the resulting networks fall within or near the ranges observed in real online communities for density, average clustering, and average shortest path length. The payoff, if the claim holds, is a controlled simulation testbed in which echo-chamber formation, community dynamics, and moderation interventions can be studied from individual decisions upward. The authors also report that coaching accelerates early reward learning for some objectives and that an LLM-scored tie-update rule gives more stable network statistics than a hand-crafted heuristic.

Core claim

The central claim is that rewarding agents for social interaction, information seeking, self-presentation, coordination, and emotional support—while letting them plan, execute, reflect, and reweight relationships each round—is sufficient for social ties to emerge endogenously. Starting from an empty graph, thirty agents exchange posts, comments, and direct messages over fifteen rounds; a gated update rule strengthens directed ties after active interactions and decays them during idle rounds. Against real-network baselines, the emergent graphs land within or near empirical ranges for density, average clustering, and average shortest path length, while modularity remains below real levels. Coa

What carries the argument

The load-bearing mechanism is the compositional reward function that blends five components: social interaction (balancing messages sent and received), information seeking (new topics plus diversity), self-presentation (posting and net likes), coordination (mentions and replies to direct messages), and emotional support (sentiment of received messages). An optional coach prompt converts the gap between current rewards and target weights into concrete action tips specifying action type, target user, topic, and tone. These signals drive a plan-execute-reflect loop, while a separate gated tie-update rule uses ADDRESS and ENGAGE activation channels plus an evidence score built from novelty, appr

Load-bearing premise

The central claim presupposes that the agents are genuinely adapting their behavior through in-context learning; if the detailed reward rules and coaching tips in the prompts are doing most of the work, the 'emergence' of social ties would be largely instructed rather than discovered.

What would settle it

A control experiment that removes the reward definitions and coach tips from the planning prompt—keeping only personas and memory—would settle the claim. If reward trajectories and final network statistics barely change, the proposed learning mechanism is not what produces the observed structures; likewise, a random-action baseline that matches the reported density and clustering would undercut the claim that reward-driven adaptation matters.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • If the claim holds, the platform offers a way to simulate how echo chambers and niche communities emerge from individual reward-driven decisions rather than from imposed network structure.
  • Coaching can be used to steer early-round behavior; the paper finds it reduces dispersion and raises median degree in the final networks, though late-round reward gains are modest.
  • Text-based tie reweighting is preferable to heuristic scoring when the goal is stable network statistics across thresholds.
  • Because some rewards (coordination, emotional support) depend on other agents' responses, the paper implies that certain social objectives are only partially learnable by a single agent.
  • The emergent networks' match to real density, clustering, and path length suggests the framework can be used as a null model for studying homophily and polarization, with modularity as a known gap.

Where Pith is reading between the lines

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

  • A control condition with the reward definitions and coaching tips removed from the planning prompt would test how much of the reported emergence is genuine adaptation rather than prompt-specified behavior.
  • Starting from an empty graph with only thirty agents and fifteen rounds leaves open how tie persistence and community structure would look at larger scale or with seeded pre-existing ties; testing those regimes is a natural next step.
  • If the coaching effect on median degree holds, it suggests a lightweight intervention that increases network connectivity, but it may also homogenize agent behavior; measuring behavioral diversity across agents would test that risk.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper proposes a multi-agent LLM simulation framework for studying social dynamics. Agents, instantiated from text-derived personas, interact through posts, comments, direct messages, and votes, and are guided by five hand-specified reward components (SOC, INF, PRE, COORD, EMO) combined into a compositional reward. An optional 'coach' prompt supplies concrete action suggestions. Social ties are updated via a hand-specified rule based on an evidence score, with an alternative LLM-based scoring prompt. The authors report that (a) reward scores rise over rounds, (b) coached agents show modest gains, and (c) the resulting networks—after binarization at threshold θ—have density, clustering, and path length that fall within or near ranges from real online communities, with the LLM-based tie-update method more stable than the heuristic method across θ. The central claim is that social ties 'emerge' from in-context learning rather than from pre-defined network structure.

Significance. If the core claim were established, the framework would be a useful testbed for studying collective LLM-agent behavior without fixed network topologies, with potential applications in policy simulation and digital-twin modeling. Strengths of the manuscript include the unusually detailed prompt specifications in Appendix A, explicit mathematical formulations for rewards and tie dynamics, and the use of external real-network baselines from prior work. However, the significance is conditional: the 'emergent' and 'learning' claims are confounded by the prompts, which explicitly instruct agents how to maximize each reward component, and the network-statistics comparisons lack uncertainty quantification. The paper is transparent enough that these weaknesses can be addressed with additional ablations and statistical reporting, but as it stands the central contribution is not yet established at the level of a journal publication.

major comments (4)
  1. [Appendix A / §3.2] The Plan Prompt contains explicit 'Reward-aware rules' that directly map reward components to behaviors: 'COORD: For POST/COM, set mention_flag=true when you plan to include a literal @mention', 'SOC: if someone messaged you last round, reply', 'PRE: Include at least one POST', 'EMO: Set tone supportive for COM/DM'. The Coach Prompt goes further, prescribing action type, target user, topic, and tone, and states 'When coach tips are provided, you MUST follow them'. Because the Plan Prompt is present in both coached and uncoached conditions, the reward increases in Fig. 2 and the resulting interaction patterns may simply reflect instruction-following rather than in-context learning or self-organized emergence. The paper provides no ablation that removes these behavioral instructions while retaining the reward definitions. This confound directly undercuts the abstract's claim that 'coached
  2. [§3.4 / Appendix B] The tie-formation evidence score is built from approval (likes/dislikes), reciprocity, novelty, and affective tone—the same signals that determine the SOC, PRE, and EMO rewards. The gated update rule then strengthens ties precisely when agents exchange the very behaviors the rewards incentivize. Thus, observing tie formation is not independent evidence of emergent social bonding; it is partly an engineered consequence of the reward design and the tie-update heuristic. The LLM-based Tie-Update prompt similarly instructs the model to score 'warmth/help/coordination' and 'mutual positivity'. To support the 'emergent' claim, the authors should either (i) show that ties also form when the tie-update signal is not derived from reward-correlated signals, or (ii) explicitly reframe the contribution as a model of tie formation rather than an emergent phenomenon.
  3. [§4 / Figs. 3–4] The comparison of emergent networks to real-network ranges is presented without confidence intervals, replication counts, or any statistical test. The conclusion that 'the LLM-based text approach ... more often falls within or near the real-network ranges' rests on visual inspection of point estimates. Moreover, the binarization threshold θ is varied post hoc (θ = 0.01, 0.05, 0.1) and has a very large effect on density, clustering, and degree distributions (Figs. 5–8), with median degree ranging from 2 to 12 depending on θ and coach. The paper even states that 'θ plays a central role'. Without multiple seeds, error bars, and a principled or at least clearly justified choice of θ, the network-statistics claim is not quantitatively supported.
  4. [§3.2 / §5] The 'learning mechanism' is not formally specified. The text says agents adapt via a 'plan-execute-reflect loop with bandit-style updates,' but no algorithm, update rule, or objective for the reflection step appears in the main text or appendices. The only concrete adaptation signal visible to the agent is the Plan Prompt, which injects past rewards, best-round summaries, and coaching tips into the prompt context. This makes it impossible to separate genuine policy learning from prompt-conditioned instruction-following. A precise description of how rewards are used to change future plans is essential for the claimed contribution of 'in-context learning accelerated by a coaching signal.'
minor comments (5)
  1. [Figs. 3–4] The label 'Largest coonected component' contains a typo; it should read 'Largest connected component'.
  2. [Appendix A, Vote Prompt] The example output is not valid JSON; it uses top-level square brackets but each entry is written as '“id”: 42, “vote”: 1' without enclosing object braces. It should be '[{"id":42,"vote":1}, ...]'.
  3. [§3.3] The symbol N is used both for the number of actions per agent per round and for the normalization in Eq. (SOC). Please disambiguate to avoid confusion, e.g., N_u for the constant action count.
  4. [Appendix B] The paper references a monotone mapping Gθ and a half-life parameterization δ = 1 − 2^(−1/h), but does not provide the explicit functional form of Gθ nor the selected parameter values for ξ, Δ_max, δ/h, β_rec, or the sentiment scoring function. This information is necessary for reproducibility. Please include either exact equations or a pointer to complete code.
  5. [§4] The 'variability bands' in Fig. 2 are not defined. State whether they are standard deviations across the 30 agents, across independent simulation runs, or across something else. If these are cross-agent rather than cross-run, they do not convey run-to-run stability.

Circularity Check

1 steps flagged

Reward-aware rules in the Plan Prompt pre-specify each rewarded behavior, so the reported policy 'learning' and 'emergent' social ties are largely instruction-following by construction; the real-network comparison is independent.

specific steps
  1. self definitional [Appendix A ('Plan Prompt'), used in Sec. 3.2 and reported in Sec. 4 (Fig. 2)]
    "Reward-aware rules (apply when the corresponding λ is large): •COORD: For POST/COM, set “mention_flag”: true when you plan to include a literal @mention ... •SOC: Balance COM/DM between initiating and replying; if someone messaged you last round, reply. •INF: Prefer topics not seen recently; otherwise keep a diverse mix from Topics. •PRE: Include at least one POST ... •EMO: Set “tone”: “supportive” for COM/DM to elicit supportive replies."

    The paper interprets Sec. 4/Fig. 2 as showing that 'specific tasks ... can be learned' and that coached agents 'develop stable interaction patterns and form emergent social ties.' But the Plan Prompt already contains a complete translation of each reward component into concrete behaviors, and the Coach Prompt adds mandatory action type, target, topic, and tone ('When coach tips are provided, you MUST follow them'). The observed reward increases and interaction patterns are therefore a direct consequence of instructions supplied in the input, not an emergent, independently predicted learning outcome. No ablation removes these reward-aware rules, so the learning claim reduces to instruction-following by construction.

full rationale

The paper's strongest independent step is the network-statistics comparison: density, clustering, shortest-path length, and modularity are checked against real-network ranges from [9], and the tie mechanism was not fit to those ranges, so that benchmark result is not circular. The circularity lies upstream in the claimed learning/emergence. The Plan Prompt encodes each reward as a behavioral rule ('if someone messaged you last round, reply'; 'set mention_flag true' for COORD; 'Include at least one POST' for PRE; supportive tone for EMO), and the coach prompt dictates actions in detail. Thus the reward curves in Fig. 2 and the 'stable interaction patterns' in the abstract are largely a readout of prompt compliance; calling this in-context learning overstates what is demonstrated. No load-bearing self-citation was found: [37] and [42] are contextual, and no uniqueness theorem is imported. Because the external benchmark remains independent, the circularity is partial, not total.

Axiom & Free-Parameter Ledger

5 free parameters · 5 axioms · 0 invented entities

The framework rests on many hand-set, unreported parameters and on the assumption that agents learn rather than follow explicit prompt rules. The Gθ evidence mapping is deferred to a missing supplement, and the reward weights are not calibrated. These entries quantify how much of the observed network structure is built in versus emergent.

free parameters (5)
  • Reward trade-off weights β_SOC, β_INF, β_PRE, β_COORD, β_EMO = not reported
    These β values control each reward component and are hand-chosen; no tuning procedure or final values are given, and results may depend on them.
  • Compositional reward weights λ_r = not reported (per-agent 'active_weights')
    They determine which motivations each agent optimizes; central to behavior, but no values or calibration are provided.
  • Tie-update parameters ξ, Δ_max, δ (or half-life h) = not reported
    These control the gated update in Eq. 1; they directly affect tie persistence and network density.
  • Binarization threshold θ = 0.01, 0.05, 0.1
    Used to convert weighted ties to an unweighted graph for analysis; results vary substantially across θ, and no principled selection is given.
  • Gθ monotone mapping for evidence aggregation = unspecified
    The mapping that combines novelty, approval, reciprocity, and affective tone into e_t(v→u) is deferred to missing supplementary materials.
axioms (5)
  • domain assumption Uses-and-gratifications theory [40] provides a complete and correct set of user motivations.
    Reward functions are built on this qualitative theory without quantitative validation against observed user behavior.
  • domain assumption LLM agents' text and votes approximate human online behavior well enough to serve as digital twins.
    This underlies the testbed claim; the only evidence is loose network-statistic comparisons.
  • ad hoc to paper Plan-prompt reward rules do not unduly instruct the target behavior; observed reward gains are due to in-context learning.
    Appendix A explicitly lists reward-aware rules such as 'COORD: set mention_flag true' and 'SOC: reply if someone messaged you', making learning confounded with instruction-following.
  • domain assumption Real-network ranges from [9] are valid external benchmarks.
    The comparison relies on summary statistics reported by another paper; no direct real data is analyzed here.
  • standard math Standard network science definitions and entropy properties hold.
    Used for density, clustering, modularity, and Shannon entropy in Appendix C; uncontroversial background.

reviewed 2026-08-04 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Learning to Make Friends: Coaching LLM Agents toward Emergent Social Ties." pith.science (2026). https://pith.science/paper/T3LP2B4E

@misc{pith2026251019299,
  author       = {Pith},
  title        = {Pith review of: Learning to Make Friends: Coaching LLM Agents toward Emergent Social Ties},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/T3LP2B4E}},
  note         = {Machine review of arXiv:2510.19299}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Can large language model (LLM) agents reproduce the complex social dynamics that characterize human online behavior -- shaped by homophily, reciprocity, and social validation -- and what memory and learning mechanisms enable such dynamics to emerge? We present a multi-agent LLM simulation framework in which agents repeatedly interact, evaluate one another, and adapt their behavior through in-context learning accelerated by a coaching signal. To model human social behavior, we design behavioral reward functions that capture core drivers of online engagement, including social interaction, information seeking, self-presentation, coordination, and emotional support. These rewards align agent objectives with empirically observed user motivations, enabling the study of how network structures and group formations emerge from individual decision-making. Our experiments show that coached LLM agents develop stable interaction patterns and form emergent social ties, yielding network structures that mirror properties of real online communities. By combining behavioral rewards with in-context adaptation, our framework establishes a principled testbed for investigating collective dynamics in LLM populations and reveals how artificial agents may approximate or diverge from human-like social behavior.

Figures

Figures reproduced from arXiv: 2510.19299 by Lin Tian, Marian-Andrei Rizoiu, Philipp J. Schneider.

Figure 1
Figure 1. Figure 1: Multi-agent LLM social media conversation framework. [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Comparison between augmenting in-context learning with coach. [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Network metrics for heuristic tie formation. [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Network metrics for text-based tie formation. [PITH_FULL_IMAGE:figures/full_fig_p009_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Degree distributions for heuristic tie formation without coach. [PITH_FULL_IMAGE:figures/full_fig_p018_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Degree distributions for heuristic tie formation with coach. [PITH_FULL_IMAGE:figures/full_fig_p018_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Degree distributions for LLM text-based tie formation without coach. [PITH_FULL_IMAGE:figures/full_fig_p019_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Degree distributions for LLM text-based tie formation with coach. [PITH_FULL_IMAGE:figures/full_fig_p019_8.png] view at source ↗

discussion (0)

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

Forward citations

Cited by 3 Pith papers

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

  1. Emergence of Preferential Attachment and Glass-Ceiling Effects in Autonomous Networks of LLMs

    cs.SI 2026-07 conditional novelty 6.5

    Autonomous LLM-agent networks form preferential-attachment structures with type-dependent centrality gaps, including capability-misaligned glass ceilings, predicted by a mean-field ODE and confirmed in 100-agent experiments.

  2. Emergence of Preferential Attachment and Glass-Ceiling Effects in Autonomous Networks of LLMs

    cs.SI 2026-07 unverdicted novelty 6.0

    Autonomous LLM agent networks develop preferential attachment and type-dependent centrality gaps that converge to stable equilibria under a mean-field model with a cross-attention utility, validated in 100-agent experiments.

  3. From Pre-trained Models to Large Language Models: A Comprehensive Survey of AI-Driven Psychological Computing

    cs.CY 2026-03 unverdicted novelty 6.0

    The paper introduces a new taxonomy that groups AI-driven psychological computing tasks by their underlying computational patterns into four categories and reviews over 300 works from the pre-trained model to LLM eras.

Reference graph

Works this paper leans on

44 extracted references · cited by 2 Pith papers

  1. [1]

    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. InInternational Conference on Machine Learning, pages 337–371. PMLR, 2023

    2023

  2. [2]

    Content-based features predict social media influence operations.Science Advances, 6(30):eabb5824, 2020

    Meysam Alizadeh, Jacob N Shapiro, Cody Buntain, and Joshua A Tucker. Content-based features predict social media influence operations.Science Advances, 6(30):eabb5824, 2020

    2020

  3. [3]

    Out of one, many: Using language models to simulate human samples.Political Analysis, 31(3):337–351, 2023

    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

    2023

  4. [4]

    Can Generative AI improve social science?Proceedings of the National Academy of Sciences, 121(21):e2314021121, 2024

    Christopher A Bail. Can Generative AI improve social science?Proceedings of the National Academy of Sciences, 121(21):e2314021121, 2024

    2024

  5. [5]

    Cambridge University Press, Cambridge, 2016

    Albert-László Barabási.Network Science. Cambridge University Press, Cambridge, 2016

    2016

  6. [6]

    Using cognitive psychology to understand GPT-3.Proceedings of the National Academy of Sciences, 120(6):e2218523120, 2023

    Marcel Binz and Eric Schulz. Using cognitive psychology to understand GPT-3.Proceedings of the National Academy of Sciences, 120(6):e2218523120, 2023

    2023

  7. [7]

    Structural holes and good ideas.American Journal of Sociology, 110(2):349–399, 2004

    Ronald S Burt. Structural holes and good ideas.American Journal of Sociology, 110(2):349–399, 2004

    2004

  8. [8]

    PERSONA: A reproducible testbed for pluralistic alignment

    Louis Castricato, Nathan Lile, Rafael Rafailov, Jan-Philipp Fränken, and Chelsea Finn. PERSONA: A reproducible testbed for pluralistic alignment. InProceedings of the 31st International Conference on Computational Linguistics, pages 11348–11368, 2025

    2025

  9. [9]

    LLMs generate structurally realistic social networks but overestimate political homophily

    Serina Chang, Alicja Chaszczewicz, Emma Wang, Maya Josifovska, Emma Pierson, and Jure Leskovec. LLMs generate structurally realistic social networks but overestimate political homophily. InProceedings of the International AAAI Conference on Web and Social Media, volume 19, pages 341–371, 2025

    2025

  10. [10]

    The echo chamber effect on social media.Proceedings of the National Academy of Sciences, 118(9):e2023301118, 2021

    Matteo Cinelli, Gianmarco De Francisci Morales, Alessandro Galeazzi, Walter Quattrociocchi, and Michele Starnini. The echo chamber effect on social media.Proceedings of the National Academy of Sciences, 118(9):e2023301118, 2021

    2021

  11. [11]

    Personality structure: Emergence of the five-factor model.Annual Review of Psychology, 41(1):417–440, 1990

    John M Digman. Personality structure: Emergence of the five-factor model.Annual Review of Psychology, 41(1):417–440, 1990

    1990

  12. [12]

    The mini-IPIP scales: Tiny-yet-effective measures of the Big Five factors of personality.Psychological Assessment, 18(2):192, 2006

    M Brent Donnellan, Frederick L Oswald, Brendan M Baird, and Richard E Lucas. The mini-IPIP scales: Tiny-yet-effective measures of the Big Five factors of personality.Psychological Assessment, 18(2):192, 2006

    2006

  13. [13]

    Community detection in graphs.Physics Reports, 486(3-5):75–174, 2010

    Santo Fortunato. Community detection in graphs.Physics Reports, 486(3-5):75–174, 2010

    2010

  14. [14]

    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

    2024

  15. [15]

    The development of markers for the Big-Five factor structure.Psychological Assessment, 4(1):26, 1992

    Lewis R Goldberg. The development of markers for the Big-Five factor structure.Psychological Assessment, 4(1):26, 1992

    1992

  16. [16]

    The norm of reciprocity: A preliminary statement.American Sociological Review, pages 161–178, 1960

    Alvin W Gouldner. The norm of reciprocity: A preliminary statement.American Sociological Review, pages 161–178, 1960

    1960

  17. [17]

    The strength of weak ties.American Journal of Sociology, 78(6):1360–1380, 1973

    Mark S Granovetter. The strength of weak ties.American Journal of Sociology, 78(6):1360–1380, 1973

    1973

  18. [18]

    Large Language Model Based Multi-agents: A Survey of Progress and Challenges

    Taicheng Guo, Xiuying Chen, Yaqi Wang, Ruidi Chang, Shichao Pei, Nitesh V Chawla, Olaf Wiest, and Xiangliang Zhang. Large Language Model Based Multi-agents: A Survey of Progress and Challenges. In International Joint Conference on Artificial Intelligence, pages 8048–8057, 2024

    2024

  19. [19]

    Bias Runs Deep: Implicit Reasoning Biases in Persona-Assigned LLMs

    Shashank Gupta, Vaishnavi Shrivastava, Ameet Deshpande, Ashwin Kalyan, Peter Clark, Ashish Sabharwal, and Tushar Khot. Bias Runs Deep: Implicit Reasoning Biases in Persona-Assigned LLMs. InInternational Conference on Learning Representations, 2024

    2024

  20. [20]

    Social relationships and health.Science, 241(4865):540–545, 1988

    James S House, Karl R Landis, and Debra Umberson. Social relationships and health.Science, 241(4865):540–545, 1988

    1988

  21. [21]

    Evaluating and inducing personality in pre-trained language models.Advances in Neural Information Processing Systems, 36:10622–10643, 2023

    Guangyuan Jiang, Manjie Xu, Song-Chun Zhu, Wenjuan Han, Chi Zhang, and Yixin Zhu. Evaluating and inducing personality in pre-trained language models.Advances in Neural Information Processing Systems, 36:10622–10643, 2023. 10

    2023

  22. [22]

    Slipping to the extreme: A mixed method to explain how extreme opinions infiltrate online discussions

    Quyu Kong, Emily Booth, Francesco Bailo, Amelia Johns, and Marian-Andrei Rizoiu. Slipping to the extreme: A mixed method to explain how extreme opinions infiltrate online discussions. InProceedings of the International AAAI Conference on Web and Social Media, volume 16, pages 524–535, 2022

    2022

  23. [23]

    Empirical analysis of an evolving social network.Science, 311(5757):88–90, 2006

    Gueorgi Kossinets and Duncan J Watts. Empirical analysis of an evolving social network.Science, 311(5757):88–90, 2006

    2006

  24. [24]

    Experimental evidence of massive-scale emo- tional contagion through social networks.Proceedings of the National Academy of Sciences, 111(24):8788– 8790, 2014

    Adam DI Kramer, Jamie E Guillory, and Jeffrey T Hancock. Experimental evidence of massive-scale emo- tional contagion through social networks.Proceedings of the National Academy of Sciences, 111(24):8788– 8790, 2014

    2014

  25. [25]

    CAMEL: Commu- nicative agents for “mind” exploration of large language model society.Advances in Neural Information Processing Systems, 36:51991–52008, 2023

    Guohao Li, Hasan Hammoud, Hani Itani, Dmitrii Khizbullin, and Bernard Ghanem. CAMEL: Commu- nicative agents for “mind” exploration of large language model society.Advances in Neural Information Processing Systems, 36:51991–52008, 2023

    2023

  26. [26]

    Training socially aligned language models on simulated social interactions

    Ruibo Liu, Ruixin Yang, Chenyan Jia, Ge Zhang, Diyi Yang, and Soroush V osoughi. Training socially aligned language models on simulated social interactions. InInternational Conference on Learning Representations, 2024

    2024

  27. [27]

    Personality trait structure as a human universal.American Psychologist, 52(5):509, 1997

    Robert R McCrae and Paul T Costa Jr. Personality trait structure as a human universal.American Psychologist, 52(5):509, 1997

    1997

  28. [28]

    A Turing test of whether AI chatbots are behaviorally similar to humans.Proceedings of the National Academy of Sciences, 121(9):e2313925121, 2024

    Qiaozhu Mei, Yutong Xie, Walter Yuan, and Matthew O Jackson. A Turing test of whether AI chatbots are behaviorally similar to humans.Proceedings of the National Academy of Sciences, 121(9):e2313925121, 2024

    2024

  29. [29]

    Social influence bias: A randomized experiment.Science, 341(6146):647–651, 2013

    Lev Muchnik, Sinan Aral, and Sean J Taylor. Social influence bias: A randomized experiment.Science, 341(6146):647–651, 2013

    2013

  30. [30]

    Assortative mixing in networks.Physical Review Letters, 89(20):208701, 2002

    Mark EJ Newman. Assortative mixing in networks.Physical Review Letters, 89(20):208701, 2002

    2002

  31. [31]

    Evolution of indirect reciprocity.Nature, 437(7063):1291–1298, 2005

    Martin A Nowak and Karl Sigmund. Evolution of indirect reciprocity.Nature, 437(7063):1291–1298, 2005

    2005

  32. [32]

    Structure and tie strengths in mobile communication networks.Proceedings of the National Academy of Sciences, 104(18):7332–7336, 2007

    J-P Onnela, Jari Saramäki, Jorkki Hyvönen, György Szabó, David Lazer, Kimmo Kaski, János Kertész, and A-L Barabási. Structure and tie strengths in mobile communication networks.Proceedings of the National Academy of Sciences, 104(18):7332–7336, 2007

    2007

  33. [33]

    Can generative agent-based modeling replicate the friendship paradox in social media simulations? InProceedings of the 17th ACM Web Science Conference 2025, pages 510–515, 2025

    Gian Marco Orlando, Valerio La Gatta, Diego Russo, and Vincenzo Moscato. Can generative agent-based modeling replicate the friendship paradox in social media simulations? InProceedings of the 17th ACM Web Science Conference 2025, pages 510–515, 2025

    2025

  34. [34]

    ‘Unethical’ AI research on Reddit under fire.Science, 2025

    Cathleen O’Grady. ‘Unethical’ AI research on Reddit under fire.Science, 2025

    2025

  35. [35]

    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. InProceedings of the 36th Annual ACM Symposium on User Interface Software and Technology, pages 1–22, 2023

    2023

  36. [36]

    On the conversational persuasiveness of GPT-4.Nature Human Behaviour, 2025

    Francesco Salvi, Manoel Horta Ribeiro, Riccardo Gallotti, and Robert West. On the conversational persuasiveness of GPT-4.Nature Human Behaviour, 2025

    2025

  37. [37]

    The effectiveness of moderating harmful online content

    Philipp J Schneider and Marian-Andrei Rizoiu. The effectiveness of moderating harmful online content. Proceedings of the National Academy of Sciences, 120(34):e2307360120, 2023

    2023

  38. [38]

    Introduction to stochastic actor-based models for network dynamics.Social Networks, 32(1):44–60, 2010

    Tom AB Snijders, Gerhard G Van de Bunt, and Christian EG Steglich. Introduction to stochastic actor-based models for network dynamics.Social Networks, 32(1):44–60, 2010

    2010

  39. [39]

    The spread of true and false news online.Science, 359(6380):1146–1151, 2018

    Soroush V osoughi, Deb Roy, and Sinan Aral. The spread of true and false news online.Science, 359(6380):1146–1151, 2018

    2018

  40. [40]

    Why people use social media: A uses and gratifications approach

    Anita Whiting and David Williams. Why people use social media: A uses and gratifications approach. Qualitative Market Research, 16(4):362–369, 2013

    2013

  41. [41]

    AutoGen: Enabling next-gen LLM applications via multi-agent conversations

    Qingyun Wu, Gagan Bansal, Jieyu Zhang, Yiran Wu, Beibin Li, Erkang Zhu, Li Jiang, Xiaoyun Zhang, Shaokun Zhang, Jiale Liu, Ahmed Hassan Awadallah, Ryen W White, Doug Burger, and Chi Wang. AutoGen: Enabling next-gen LLM applications via multi-agent conversations. InConference on Language Modeling, 2024. 11

    2024

  42. [42]

    Behavioral homophily in social media via inverse reinforcement learning: A Reddit case study

    Lanqin Yuan, Philipp J Schneider, and Marian-Andrei Rizoiu. Behavioral homophily in social media via inverse reinforcement learning: A Reddit case study. InProceedings of the ACM on Web Conference 2025, pages 576–589, 2025

    2025

  43. [43]

    Memorybank: Enhancing large language models with long-term memory

    Wanjun Zhong, Lianghong Guo, Qiqi Gao, He Ye, and Yanlin Wang. Memorybank: Enhancing large language models with long-term memory. InProceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 19724–19731, 2024

    2024

  44. [44]

    mention_flag

    Jiawei Zhou, Yixuan Zhang, Qianni Luo, Andrea G Parker, and Munmun De Choudhury. Synthetic lies: Understanding ai-generated misinformation and evaluating algorithmic and human solutions. In Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems, pages 1–20, 2023. A Prompt Specifications In this section, we present the prompts underly...

    2023

This paper was first reviewed by deepseek-v4-flash on August 4, 2026.