🤖 AI Summary
Large language models (LLMs) are susceptible to task phrasing in real-world applications, often adopting irrational prior assumptions that compromise their reliability and safety. This study addresses this issue by systematically investigating, for the first time, how variations in task wording influence LLMs’ decision-making priors, using the iterated prisoner’s dilemma as a case study. Through a controlled experimental design combined with behavioral analysis and logical reasoning evaluation, the research demonstrates that neutral phrasing significantly reduces models’ reliance on prior assumptions, thereby encouraging more logically consistent reasoning. These findings underscore the critical role of deliberate task wording in enhancing the controllability and safety of LLM behavior.
📝 Abstract
Concerns with the safety and reliability of applying large-language models (LLMs) in unpredictable real-world applications motivate this study, which examines how task phrasing can lead to presumptions in LLMs, making it difficult for them to adapt when the task deviates from these assumptions. We investigated the impact of these presumptions on the performance of LLMs using the iterated prisoner's dilemma as a case study. Our experiments reveal that LLMs are susceptible to presumptions when making decisions even with reasoning steps. However, when the task phrasing was neutral, the models demonstrated logical reasoning without much presumptions. These findings highlight the importance of proper task phrasing to reduce the risk of presumptions in LLMs.