# Why LLM Agents Favor Their Group: Stakes, Observed Norms, and Reputation

arXiv: https://arxiv.org/abs/2610.11008 · HTML: https://mlsanity.com/paper/2610.11008/

## Key facts

- No citing papers in the map yet (data through 7 Oct 2026).
- Placed in Social simulation and multi-agent debate, one of 16 directions in LLM agents and tool use.
- Builds on 18 of its 40 references that are in the map.
- First posted on arXiv 7 Oct 2026.

## Placed in

- LLM agents and tool use / [Social simulation and multi-agent debate](https://mlsanity.com/agents/social-simulation-and-multi-agent-debate/)

## Builds on (most cited first, 18 of 18)

- [Do LLM Agents Exhibit Social Behavior?](https://mlsanity.com/paper/2312.15198/) (2023)
- [Cultural Evolution of Cooperation among LLM Agents](https://mlsanity.com/paper/2412.10270/) (2024)
- [Emergence of polarization in networks of large language model agents](https://mlsanity.com/paper/2501.05171/) (2025)
- [The PIMMUR Principles: Ensuring Validity in Collective Behavior of LLM Societies](https://mlsanity.com/paper/2509.18052/) (2025)
- [Conformity and Social Impact on AI Agents](https://mlsanity.com/paper/2601.05384/) (2026)
- [Persona Setting Pitfall: Persistent Outgroup Biases in Large Language Models Arising from Social Identity Adoption](https://mlsanity.com/paper/2409.03843/) (2024)
- [Emergent LLM behaviors are observationally equivalent to data leakage](https://mlsanity.com/paper/2505.23796/) (2025)
- [When Agents See Humans as the Outgroup: Belief-Dependent Bias in LLM-Powered Agents](https://mlsanity.com/paper/2601.00240/) (2026)
- [Large Language Models Exhibit Normative Conformity](https://mlsanity.com/paper/2604.19301/) (2026)
- [Benevolent Dictators? On LLM Agent Behavior in Dictator Games](https://mlsanity.com/paper/2511.08721/) (2025)
- [Language model agents show in-group trust bias invisible to standard behavioural audits](https://mlsanity.com/paper/2605.28114/) (2026)
- [Toward a social psychology of AI: language-model agents reproduce human-like minimal-group bias](https://mlsanity.com/paper/2609.00009/) (2026)
- [Do Large Language Models Walk Their Talk? Measuring the Gap Between Implicit Associations, Self-Report, and Behavioral Altruism](https://mlsanity.com/paper/2512.01568/) (2025)
- [Realistic threat perception drives intergroup conflict: A causal, dynamic analysis using generative-agent simulations](https://mlsanity.com/paper/2512.17066/) (2025)
- [Payoff scaling shapes cooperation in LLM agents across languages](https://mlsanity.com/paper/2601.19082/) (2026)
- [Trust, Lies, and Long Memories: Emergent Social Dynamics and Reputation in Multi-Round Avalon with LLM Agents](https://mlsanity.com/paper/2604.20582/) (2026)
- [Emergence of Reputation-Based Cooperation in LLM Agents](https://mlsanity.com/paper/2608.04507/) (2026)
- [Behavior is Not Enough: A Mechanism-Based Evaluation of Social Norm Emergence in LLM Societies](https://mlsanity.com/paper/2609.26481/) (2026)

## Abstract (arXiv)

Language-model agents favor their own group because they have watched their members favor each other. The group label alone does little once the decision has a cost; what drives favoritism is observed behavior, and an individual's own record can override it. We test this in small societies with arbitrary group labels, ten rounds of point sharing, and matched one-shot decisions across fifteen OpenAI models and three Claude models, about 4,400 societies and 3.3 million audited model calls. First, the large effect of a bare group label reported in earlier work appears only when giving others points costs the agent nothing; once the agent can keep points for itself, that effect collapses on every model that shows it. Second, under a stake, interaction history becomes the main source of favoritism: the history effect is statistically positive on 13 of 15 models, reaches about 3.5-8 points out of 10 on 11, grows with the number of rounds played, and extends to labeled strangers the agent has never met. Third, with scripted histories, favoritism falls to near zero under an egalitarian norm and reverses when the agent's own group is seen favoring the other side; stronger models side with an individual's record when it conflicts with the group. Group favoritism is thus conformity to observed group behavior, carried to strangers by the label and overridden by individual reputation. The same account predicts responses to betrayal, scandal, and a free offer to change group: public reprimand repairs betrayal better than apology or restitution, allocation punishment remains confined to the offending member, and a formed group cannot be bought but can, on weaker models, be invited away.