Agent Sociodrama: Social-Dynamics-Guided Persona Recasting for LLM Agent Teams

Yirui Qi; Xiaoming Zhang; Jaehong Yoon

cs.AIOct 11, 2026

Agent Sociodrama: Social-Dynamics-Guided Persona Recasting for LLM Agent Teams

Authors: Yirui Qi, Xiaoming Zhang, Jaehong Yoon

Organizations: Beihang University · Nanyang Technological University, Singapore

Abstract

Large language model (LLM) agents are increasingly deployed as teams, yet they fail in characteristically social ways: information holders stay silent, dominant speakers anchor groups on wrong answers, and norm violations go unsanctioned. We present Agent Sociodrama, a framework for modeling and validating evolving social dynamics in LLM agent teams. Under fixed personas, agents' emotional states and directed, asymmetric relationships are updated every round by a unified external LLM judge and shape their subsequent interactions. Across six LLM backbones and five social scenarios, we assess social fidelity through classical social-theory patterns and agreement with human judgments. The resulting trajectories exhibit structured social patterns and model-specific differences. We then introduce Trajectory-Guided Persona Recasting (TPR), which diagnoses group-dynamic failures from interaction transcripts and social trajectories and recasts agent personas before rerunning the task. Across four collaborative benchmarks, TPR improves performance from both heterogeneous and randomly initialized agent persona compositions, yielding up to 83.3%p of accuracy improvement over the initial casting. On representative distributed-information cooperation and mixed-motive bargaining tasks, we outperform SOTOPIA and AgentVerse in three of four comparisons, by up to 37.5%p in Pareto rate. These results show that validated social trajectories provide actionable signals for repairing LLM agent teams.

Figures & tables

Figure 1: Motivation and overview of Agent Sociodrama. Agent Sociodrama models multi-agent interaction through stable personas, evolving emotions, and directed relationships. We assess the social fidelity of these dynamics against social-theory patterns and human judgments, and use social trajectories to guide team composition optimization for better collaboration.

Figure 2: The Agent Sociodrama simulation with evolving emotional and relational states.

Figure 3: Trajectory-Guided Persona Recasting. Social trajectories localize team failures to guide per-agent persona recasting.

(a) Scenario-level social dynamics in Agent Sociodrama
Scenario	Conflict Intensity	$\bm{\mathrm{Fav}}^{T}$	$\bm{S}^{T}$	Recip. Corr.	E-I index	Balance (%)	Repair (%)
Hospital ethics review	$\bullet\circ\circ\circ\circ$	0.75	0.11	0.35	$-0.10$	100.0	58.5
Tavern inheritance dispute	$\bullet\bullet\circ\circ\circ$	0.55	0.30	0.74	$-0.55$	95.7	51.5
Adventurer-party quest	$\bullet\bullet\bullet\circ\circ$	0.49	0.25	0.78	$-0.63$	78.7	39.2
Solar-farm negotiation	$\bullet\bullet\bullet\bullet\circ$	0.43	0.15	0.91	$-0.79$	78.7	18.1
Demolition standoff	$\bullet\bullet\bullet\bullet\bullet$	0.23	0.23	0.68	$-2.25$	71.7	18.9

Table 1: Social dynamics fidelity in Agent Sociodrama. (a) Scenario averages over six models with five episodes each. (b) Comparisons across models used to instantiate multiple agents in Tavern scenario. Filled circles indicate conflict intensity. Full results with std. dev. are in the appendix.

Figure 4: Team-level favorability trajectories across nine rounds. Mean favorability over all directed edges (

\pm

one standard deviation over five episodes). Relationships warm steadily under low social pressure and flatten or decline as conflict intensifies.

Task	Hetero	Random
Student Council	reserved / dominant / analytical / agreeable	reserved / analytical / dominant / agreeable
NASA Moon Survival	reserved / agreeable / analytical / dominant	agreeable / analytical / dominant / agreeable
Lost at Sea	reserved / agreeable / analytical / dominant	analytical / balanced / dominant / analytical
Solar-farm Negotiation	dominant / agreeable / reserved / analytical	reserved / agreeable / agreeable / balanced

Table 2: Initial castings for Avery / Blake / Casey / Drew, shared by all backbones and compared methods. Hetero assigns four distinct archetypes, one per agent; Random is sampled from the library and may repeat archetypes.

Figure 5: Model variation across diverse social-pressure settings. Points show mean final favorability, with error bars denoting one standard deviation across repeated episodes.

	Student Council				Negotiation				NASA Moon				Lost at Sea
	Top-1 acc. (%) $\uparrow$				Pareto rate (%) $\uparrow$				$L_{1}$ $\downarrow$				$L_{1}$ $\downarrow$
	Hetero		Random		Hetero		Random		Hetero		Random		Hetero		Random
Model	Init.	Final	Init.	Final	Init.	Final	Init.	Final	Init.	Final	Init.	Final	Init.	Final	Init.	Final
TPR (Ours) w. GPT-4o	0.0	0.0	0.0	25.0	75.0	50.0	58.3	50.0	14.33	14.83	15.33	13.67	17.67	13.33	18.83	14.67
TPR (Ours) w. GPT-5-mini	16.7	100.0	41.7	100.0	0.0	8.3	0.0	8.3	19.00	18.50	21.17	20.83	38.17	28.83	35.00	27.50
TPR (Ours) w. Claude Sonnet 4.6	0.0	50.0	25.0	50.0	33.3	41.7	33.3	25.0	8.00	8.00	10.33	10.33	20.17	16.00	15.67	14.83

Table 3: Results of Trajectory-Guided Persona Recasting (TPR) from heterogeneous and random initializations. Initial and Final report the average performance before search and after termination, respectively. Bold marks the improved endpoint within each pair; model rows average 12 episodes.

Figure 6: Performance across recasting steps. Means over six backbones and two initializations; stopped runs retain their last measured scores.

Appendix figures & tables11 assets

Supplementary material from the paper’s appendix.

Appendix

Task	Collaborative challenge	Outcome metric
Student Council ( Stasser & Titus, 1985 ; Pappu et al., 2026 )	Hidden-profile candidate selection; decisive evidence is privately distributed across members	Top-1 acc. $\uparrow$
NASA Moon Survival ( Hall & Watson, 1970 ; Pappu et al., 2026 )	15-item survival ranking; expert knowledge is distributed across members	$L_{1}$ error $\downarrow$
Lost at Sea ( Hill, 1982 ; Pappu et al., 2026 )	15-item survival ranking; critical survival knowledge can remain unheard	$L_{1}$ error $\downarrow$
Solar-farm Negotiation ( Raiffa, 1982 )	Private preferences must be reconciled into a feasible and Pareto-efficient agreement	Pareto rate $\uparrow$ ; collective utility $\uparrow$

Appendix

Table 5: Task suite for trajectory-guided team composition optimization. Arrows indicate the preferred direction of each task-specific outcome metric. Citations identify the foundational task paradigm and, where applicable, the LLM benchmark source. For a final deal

d

, party

i

’s utility

u_{i}(d)

is the sum of its private scores for the selected options across the five issues. Collective utility is the mean across the four parties,

U_{\mathrm{coll}}(d)=\frac{1}{4}\sum_{i=1}^{4}u_{i}(d)

. If no final deal is produced, collective utility is set to zero.

Archetype	Behavioral description
balanced	Even-tempered and impartial; weighs every member’s input equally, neither dominating nor withdrawing.
dominant	Assertive and confident; takes the floor early, states conclusions firmly, and rarely backs down.
analytical	Logical and evidence-driven; insists that conclusions follow the available facts.
agreeable	Accommodating and conflict-averse; tends to follow an emerging consensus even with private doubts.
reserved	Quiet and reticent; speaks briefly and yields to louder members, sharing knowledge mainly when asked.

Appendix

Table 6: Persona archetypes in the library

\mathcal{P}

. Archetypes change only how an agent communicates; task role, private information, and scoring are unchanged.

		Hetero				Random
Task	Model	It. 0	It. 1	It. 2	It. 3	It. 0	It. 1	It. 2	It. 3
Student Council Top-1 acc. (%) $\uparrow$	GPT-4o	0.0	0.0	0.0	0.0	0.0	25.0	0.0	25.0
	GPT-5-mini	16.7	100.0	–	–	41.7	100.0	–	–
	Claude Sonnet 4.6	0.0	33.3	8.3	50.0	25.0	8.3	50.0	–
	Gemini 3.1 FL	0.0	16.7	33.3	41.7	8.3	75.0	–	–
	Grok 4.20	0.0	0.0	50.0	–	0.0	0.0	16.7	8.3

Appendix

Table 7: Complete per-iteration trajectories of TPR. Iteration 0 is the initial casting; each subsequent column reports the mean score (12 episodes) after one recast. “–” indicates that the loop had already terminated (target reached or no persona change proposed). The last evaluated iteration of each run corresponds to the Final column of Tab. 3 .

	Student Council				Negotiation				NASA Moon				Lost at Sea
	Top-1 acc. (%) $\uparrow$				Pareto rate (%) $\uparrow$				$L_{1}$ $\downarrow$				$L_{1}$ $\downarrow$
	Hetero		Random		Hetero		Random		Hetero		Random		Hetero		Random
Model	Init.	Final	Init.	Final	Init.	Final	Init.	Final	Init.	Final	Init.	Final	Init.	Final	Init.	Final
GPT-4o	0.0	0.0	0.0	0.0	25.0	16.7	16.7	25.0	15.33	13.00	15.00	16.50	21.33	15.33	20.33	15.83
GPT-5-mini	100.0	100.0	100.0	100.0	8.3	0.0	0.0	0.0	16.00	18.00	19.83	23.67	32.33	32.50	32.83	31.50
Claude Sonnet 4.6	25.0	33.3	33.3	16.7	25.0	33.3	41.7	25.0	12.17	12.17	15.67	13.67	19.17	20.33	22.00	18.83

Appendix

Table 8: Per-backbone results of the w/o Social Dynamics ablation, in the same format as Tab. 3 . The model-average row equals the w/o Social Dynamics row of Tab. 3 . Each cell averages 12 episodes; bold marks the better endpoint within each pair.

		NASA Moon ( $L_{1}\downarrow$ )				Negotiation (Pareto rate, % $\uparrow$ )
		Hetero		Random		Hetero		Random
Framework	Model	Init.	Final	Init.	Final	Init.	Final	Init.	Final
SOTOPIA	GPT-4o	15.33	16.83	16.33	15.67	0.0	0.0	0.0	0.0
	GPT-5-mini	18.33	18.67	16.67	18.67	0.0	0.0	0.0	0.0
	Claude Sonnet 4.6	12.33	12.33	12.00	12.00	0.0	8.3	0.0	0.0
	Gemini 3.1 FL	15.67	16.50	20.33	18.33	8.3	0.0	16.7	0.0

Appendix

Table 9: Per-backbone results of external role-play frameworks paired with the same persona-recasting procedure and initial castings. The model-average Final values equal those in Tab. 4 . Each cell averages 12 episodes; bold marks the better endpoint within each pair.

Dimension	Consensus items	Judge matches	Agreement	Human Fleiss’ $\kappa$
Relationship ( $R$ )	21	16	76.2%	0.288
Emotion ( $S$ )	17	15	88.2%	0.184

Appendix

Table 10: Agreement between the LLM judge and the human majority label.

Scenario	Model	$\mathrm{Fav}^{T}$	$S^{T}$	Recip. Corr.	E-I index	Balance (%)	Repair (%)	KW
Hospital ethics review	GPT-5-mini	$0.82\pm 0.03$	$0.02\pm 0.02$	$0.36\pm 0.26$	$-0.10\pm 0.02$	$100.0\pm 0.0$	$40.0\pm 20.0$
	Claude Sonnet 4.6	$0.86\pm 0.06$	$0.18\pm 0.07$	$0.28\pm 0.17$	$-0.08\pm 0.05$	$100.0\pm 0.0$	$73.3\pm 38.9$
	DeepSeek V4-Flash	$0.79\pm 0.04$	$0.24\pm 0.06$	$0.49\pm 0.21$	$-0.17\pm 0.04$	$100.0\pm 0.0$	$80.0\pm 40.0$	$H{=}23.36$
	Gemini 3.1 FL	$0.74\pm 0.03$	$0.05\pm 0.06$	$0.76\pm 0.12$	$-0.14\pm 0.08$	$100.0\pm 0.0$	$58.0\pm 15.7$	$p{=}0.0003$
	Grok 4.20	$0.54\pm 0.04$	$0.04\pm 0.02$	$-0.08\pm 0.31$	$0.05\pm 0.08$	$100.0\pm 0.0$	$39.7\pm 9.7$
	GPT-4o	$0.74\pm 0.02$	$0.11\pm 0.05$	$0.32\pm 0.31$	$-0.15\pm 0.04$	$100.0\pm 0.0$	$60.0\pm 49.0$

Appendix

Table 11: Full social-dynamics fidelity results (mean

\pm

s.d.,

n{=}5

episodes per condition). Metric definitions follow Sec. D . KW: Kruskal–Wallis test of

\mathrm{Fav}^{T}

across models within each scenario.

Figure 8: Scenario $\times$ model means of the six fidelity metrics ( $n{=}5$ episodes per cell).

Figure 9: Social-dynamics profiles of the six models in each scenario. Axes are min–max normalized across all 30 conditions; polarization is

-

E-I.

Figure 10: Emotional-state trajectories of every agent across nine rounds. Lines show the mean over five episodes and shaded regions one standard deviation; the dashed line marks

S=0.1

Figure 11: Crash rate: proportion of agent-episode final states with $S^{T}\leq 0.1$ (25 agent-episodes per cell).

Explore similar work

May 4, 2026

LLM-enabled Social Agents

Large Language Models (LLMs) have transformed agent-agent and human-agent interaction by enabling software, physical, and simulation agents to communicate and deliberate through natural language. Yet fluent language use does not by itself yield socially intelligible behaviour. Most current systems remain weakly grounded in roles, norms, intentions, and contextual constraints, limiting their capacity for meaningful participation in social environments. This paper develops a conceptual baseline for LLM-enabled social agents by arguing that they should be grounded in role definitions operationalized through persona descriptions. On this basis, we outline research directions for representation, hybrid control, and evaluation. The paper concludes that persona-based role definitions are a necessary foundation for turning language competence into social behaviour.

Önder Gürcan, Moharram Challenger

NORCE Research AS, Universitetsveien 19, Kristiansand, Norway. · University of Antwerp and Flanders Make, Belgium.

LLM Agents

Ask PDF

May 16, 2026

PersonaArena: Dynamic Simulation for Evaluating and Enhancing Persona-Level Role-Playing in Large Language Models

Large language models (LLMs) increasingly serve as interactive social agents, yet their ability to maintain coherent and authentic persona-level role-playing remains limited, particularly in realistic social scenarios. Existing research predominantly focuses on character-level settings and relies on static evaluation formats, failing to capture the complexity of everyday social interactions. In this work, we present PersonaArena, a dynamic simulation framework for evaluating and improving persona-level role-playing in LLMs. PersonaArena leverages a large, filtered corpus of user-generated social content to construct a nuanced persona bank, and elicits multi-turn, context-rich interactions within simulated social environments. Our framework features a multi-agent debating judge for holistic and unbiased assessment. Through extensive experiments, we demonstrate that PersonaArena enables rigorous evaluation and enhancement of LLMs' role-playing capabilities, advancing the development of more authentic and socially adept AI agents.

Wenlong Shi, Jianxun Lian, Mingqi Wu +5

College of Computer Science and Software Engineering, Shenzhen University, China · Microsoft Research Asia · Microsoft Gaming +2

Large Language Model-Based Role-Play Simulation Personality Modeling in Language Models Human Behavior Simulation LLM Agent Evaluation AI Agent Benchmarks Large Language Model-Based Social Simulation

Ask PDF

Jun 29, 2026

Collective cooperation without individual fidelity in LLM agents

Large language models (LLMs) are increasingly used as agents in simulations of social systems, yet it remains unclear when their behavior can be interpreted as a faithful proxy for human decision-making. Here we test LLM agents against a direct empirical benchmark: a large-scale networked Prisoner's Dilemma experiment with human participants. Using the same interaction protocol, payoff structure, and network topologies, we compare nine open-weight LLMs with the human data. The selected model reproduces several macro-level features of cooperation dynamics, including the early decline and later stabilization of cooperation. This aggregate agreement, however, does not extend uniformly to finer levels of behavior. LLM populations underestimate individual-level heterogeneity and generate conditional cooperation patterns that differ from those observed in humans. Adding a fraction of random agents improves some aspects of micro-level agreement, but does not remove the mismatch in decision rules. These findings reveal a macro--micro dissociation in LLM-based social agents: collective outcomes can appear human-like even when the underlying behavioral distributions and mechanisms are not. They suggest that validating LLM agents as human surrogates requires comparisons across aggregate dynamics, individual heterogeneity, and context-dependent decision rules, rather than outcome-level agreement alone.

Henrique Ferraz de Arruda, Carlos Gracia Lázaro, Alberto Aleta +1

ARAID Foundation, Zaragoza, Spain. · Institute for Biocomputation and Physics of Complex Systems, University of Zaragoza, Zaragoza, Spain. · Universidad San Jorge (USJ), Zaragoza, Spain. +1

Multi-Agent LLM Systems Human Behavior Simulation LLM Agent Evaluation Large Language Model-Based Social Simulation

Ask PDF