Super-additive Cooperation in Language Model Agents

This repository contains the implementation for the paper:
"Super-additive Cooperation in Language Model Agents"
by Filippo Tonini and Lukas Galke, University of Southern Denmark.

• Paper (arXiv): arXiv:2508.15510
• Web Page: pippot.github.io/Superadditive-cooperation-LLMs

Abstract

With the prospect of autonomous artificial intelligence (AI) agents, studying their tendency for cooperative behavior becomes an increasingly relevant topic. This study is inspired by the super-additive cooperation theory, where the combined effects of repeated interactions and inter-group rivalry have been argued to be the cause for cooperative tendencies found in humans.

We devised a virtual tournament where language model agents, grouped into teams, face each other in a Prisoner's Dilemma game. By simulating both internal team dynamics and external competition, we discovered that this blend substantially boosts both overall and initial, one-shot cooperation levels.

Citation

If you find this work useful in your research, please cite the paper:

@misc{tonini2025superadditive,
      title={Super-additive Cooperation in Language Model Agents}, 
      author={Filippo Tonini and Lukas Galke},
      year={2025},
      eprint={2508.15510},
      archivePrefix={arXiv},
      primaryClass={cs.AI}
}