This repository contains the accompanying code for the paper titled "Diversified Dominance: Specialization of LLMs in Multi-Commodity Markets". The study focuses on the application and implications of Large Language Models (LLMs) in setting prices and strategies in multi-commodity markets. Specifically, it investigates whether LLMs can engage in anti-competitive behavior such as specialization to monopolize different commodities within a market.
With the increasing deployment of AI technologies in marketplaces, the study explores the behavior of LLM-based pricing agents within the frameworks of both Bertrand and Cournot competition models. Key questions addressed include:
- Are LLMs capable of engaging in anti-competitive, specialization behavior in a multi-commodity market environment?
- How do LLMs adapt their strategies in response to varying market conditions and competition?
cournot.ipynb: Jupyter notebook containing the code for the Cournot competition experiments.bertrand.ipynb: Jupyter notebook containing the code for the Bertrand competition experiments.
We would like to thank Professor Yisong Yue, Professor Peter Caradonna, and Professor Eric Mazumdar for their guidance and support throughout the project.