Simulations of normal user behavior are integral parts of cyber exercises where training and testing takes place in simulated environments. Specifically, benign user activities are essential to generate background traffic during cyber exercises and to estimate false positive rates when evaluating intrusion detection systems. Even though many user automation tools are available, developers typically only consider valid and compliant interactions with systems and applications when defining the scope of normal user behavior models. However, real legitimate users sometimes behave in ways that are non-compliant, erratic, or otherwise deviate from expected norms, and thereby generate suspicious yet benign traffic that triggers alerts from intrusion detection systems. To identify common activities in the vast space of possible user interactions and to support the design of realistic user behavior models, we assemble a list of 17 user activities that are commonly associated with false positives. We assess the relevance and frequencies of these event types with respect to their perceived priority, intent behind them, responsible actor, and circumstances in which they become noteworthy, through likert scale analysis of an expert study with 62 domain experts. Our findings reveal diverse perspectives among respondents and suggest that the behaviors leading to false positives can vary significantly between organizations.