How Personas Can Influence Agents to Play Split or Steal
Quick Answer
The study investigates how persona prompts influence strategic behavior in an iterated Split or Steal game, revealing that agents like phi4 and Ministral 3:3b consistently cooperate, while Gemma models show varied strategies.
Quick Take
Mutual Split outcomes dominated at 74%, with exploitation under 11%, indicating the significant role of model choice and persona traits in decision-making.
Key Points
- Agents were tested in 160 sessions of 15 rounds each in European Portuguese.
- Mutual Split outcomes occurred in approximately 74% of rounds.
- Models phi4 and Ministral 3:3b were consistently cooperative across temperature settings.
- Gemma3 and Gemma4 exhibited varied strategies, leading to different outcomes.
- Prosocial and Principled personas were most cooperative, while Analytical personas tended to exploit.
Paper Resources
📖 Reader Mode
~2 min readAbstract:Personas are often employed to guide large language model agents, yet their effectiveness in shaping strategic behavior in social dilemma settings remains uncertain. To address this, we examined the impact of persona prompts in an iterated Split or Steal game where persona-driven agents interacted with a Virtual Human (VH) controlled by a fixed prompt. Agents were instantiated from four open models (Ministral 3:3b, phi4:14b, Gemma3:12b, and Gemma4:e4b) at two temperature settings (0.3 and 0.7) and deterministic decision with zero temperature, while the VH was powered by GPT 4.1 mini. Across 160 sessions of 15 rounds each conducted in European Portuguese, mutual Split outcomes dominated (roughly 74 percent of rounds), with exploitation occurring in fewer than 11 percent of rounds. Model choice significantly influenced behavior: phi4 and Ministral 3:3b remained consistently cooperative across temperatures, whereas Gemma3:12b and Gemma4:e4b exhibited more varied strategies and outcomes. Analyses based on Big Five personality traits indicated that Prosocial and Principled personas were most consistently cooperative, while Analytical personas were more likely to exploit the VH. Topic analysis revealed that friendship-related dialogue aligns with Split decisions, whereas money and vengeance-related content is more prevalent in Steal outcomes; sentiment labels were predominantly neutral or happy and provided limited additional explanatory value. These findings characterize the interaction between persona prompts and model differences in repeated trust games and serve as a baseline for planned virtual reality studies involving human participants interacting with an embodied VH.
| Subjects: | Computation and Language (cs.CL); Computers and Society (cs.CY) |
| Cite as: | arXiv:2607.05398 [cs.CL] |
| (or arXiv:2607.05398v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2607.05398 arXiv-issued DOI via DataCite |
Submission history
From: Carlos Leon [view email]
[v1]
Sun, 3 May 2026 11:03:38 UTC (2,244 KB)
— Originally published at arxiv.org
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.CL
See more →TriAgent: Divergence-Aware Committees for Cost-Efficient Financial Sentiment Analysis
TriAgent introduces a cost-efficient multi-agent system for financial sentiment analysis, combining VADER, FinBERT, and Qwen2.5. It achieves an F1 score of ~0.87 with significant savings of $9.3M/year at a 10M-user scale compared to GPT-4o-mini, while also detecting hallucinations with an AUC of 0.90.