The ability to chain together sequences of information and action is pivotal to everyday acquisition of skills. Despite extensive research of sequence learning, little focus has been given to individual performance in standard tasks measuring this capability. As a result, little is known regarding what knowledge participants gain during such tasks. In the current work, an individual- and item-based analysis is performed of eye movements that occur during a spatial sequence learning task and reflect anticipation of upcoming target locations. We show that the knowledge participants acquire during the task is tightly linked to a-priori response biases they bring into the experiment. Results suggest that a-priori biases may be a sizeable influence on performance in learning experiments, that tends to be overlooked. Implications for designing and reading studies of sequence learning are discussed.