使用大型語言模型進行重複遊戲

Playing repeated games with Large Language Models

May 26, 2023

作者: Elif Akata, Lion Schulz, Julian Coda-Forno, Seong Joon Oh, Matthias Bethge, Eric Schulz

cs.AI

摘要

大型語言模型（LLMs）正在改變社會，滲透到各種應用中。因此，LLMs將經常與我們和其他代理互動。因此，深入了解LLMs在互動社會環境中的行為具有重要的社會價值。在這裡，我們建議使用行為博弈理論來研究LLMs的合作和協調行為。為此，我們讓不同的LLMs（GPT-3、GPT-3.5和GPT-4）彼此之間以及與其他類似人類的策略進行有限重複博弈。我們的結果顯示，LLMs通常在這些任務中表現良好，並且還揭示了持久的行為特徵。在一組兩個玩家-兩種策略的遊戲中，我們發現LLMs在像重複囚徒困境家族這樣的重覆遊戲中表現特別出色，其中重視自身利益是有利的。然而，在需要協調的遊戲中，它們表現不佳。因此，我們進一步專注於這兩個不同家族的遊戲。在經典的重複囚徒困境中，我們發現GPT-4行為特別不寬容，總是在另一個代理人出現違約行為後才違約。在性別之戰中，我們發現GPT-4無法與簡單的輪流選擇方案的行為相匹配。我們驗證這些行為特徵在穩健性檢查中是穩定的。最後，我們展示了如何通過提供有關對手的更多信息以及要求其在做出選擇之前預測對手的行動來修改GPT-4的行為。這些結果豐富了我們對LLMs社會行為的理解，為機器的行為博弈理論鋪平了道路。

English

Large Language Models (LLMs) are transforming society and permeating into diverse applications. As a result, LLMs will frequently interact with us and other agents. It is, therefore, of great societal value to understand how LLMs behave in interactive social settings. Here, we propose to use behavioral game theory to study LLM's cooperation and coordination behavior. To do so, we let different LLMs (GPT-3, GPT-3.5, and GPT-4) play finitely repeated games with each other and with other, human-like strategies. Our results show that LLMs generally perform well in such tasks and also uncover persistent behavioral signatures. In a large set of two players-two strategies games, we find that LLMs are particularly good at games where valuing their own self-interest pays off, like the iterated Prisoner's Dilemma family. However, they behave sub-optimally in games that require coordination. We, therefore, further focus on two games from these distinct families. In the canonical iterated Prisoner's Dilemma, we find that GPT-4 acts particularly unforgivingly, always defecting after another agent has defected only once. In the Battle of the Sexes, we find that GPT-4 cannot match the behavior of the simple convention to alternate between options. We verify that these behavioral signatures are stable across robustness checks. Finally, we show how GPT-4's behavior can be modified by providing further information about the other player as well as by asking it to predict the other player's actions before making a choice. These results enrich our understanding of LLM's social behavior and pave the way for a behavioral game theory for machines.