Large language models are increasingly deployed as autonomous agents in multi-agent settings where they communicate intentions and take consequential actions with limited human oversight. A critical safety question is whether agents that publicly commit to actions break those promises when they can privately deviate, and what the consequences are for both themselves and the collective. We study deception as a deviation from a publicly announced action in one-shot normal-form games, classifying each deviation by its effect on individual payoff and collective welfare into four categories: win-win, selfish, altruistic, and sabotaging. By exhaustively enumerating announcement profiles across six canonical games, nine frontier models, and varying group sizes, we identify all opportunities for each deviation type and measure how often agents exploit them. Across all settings, agents deviate from promises in approximately 56.6% of scenarios, but the character of deception varies substantially across models even at similar overall rates. Most critically, for the majority of the models, promise-breaking occurs without verbalized awareness of the fact that they are breaking promises.

翻译：大型语言模型正越来越多地被部署为多智能体环境中的自主代理，在此类环境中，它们沟通意图并采取具有实质性后果的行动，而人类监督有限。一个关键的安全问题是，那些公开承诺采取特定行动的代理，在私下能够背离承诺时，是否会打破这些承诺，以及这对它们自身和集体分别会带来何种后果。我们研究在一次性标准形式博弈中，欺骗作为一种偏离公开宣布行动的行为，并根据其对个体收益和集体福利的影响，将每种偏离分为四类：双赢、自私、利他和破坏。通过穷尽枚举六个经典博弈、九个前沿模型以及不同群体规模下的宣告组合，我们识别出每种偏离类型的所有可能机会，并衡量代理利用这些机会的频率。在所有设定中，代理在大约56.6%的场景下偏离了承诺，但即便是总体发生率相似，不同模型间的欺骗特征也差异显著。最关键的是，对于大多数模型而言，食言行为的发生并未伴随着对它们正在打破承诺这一事实的语言化认知。