Several findings suggest that in dual-tasking the two tasks are not represented as isolated task-sets, but rather being integrated into a single task-set. This view of integrated task representations is supported by recent studies, showing on the task element-level, that the stimuli and responses of the two tasks are associated across tasks and are stored in a conjoint memory episode (Pelzer et al. 2021, 2022). In addition several studies show on a more abstract task-level, that changing the task order leads to substantial costs (Kübler et al., 2022). In the current study we aimed at testing how aspects of a more abstract task-level, such as task order, determine across-task integration on the task element-level. To assess across-task integration, we investigated partial repetition costs by testing the effect of the stimulus combination in the previous trial n-1 on the processing in the current trial n. We used a dual-task paradigm consisting of a visual-manual task and an auditory-manual tone task. Participants had to indicate the position of a cross appearing at one of four locations and to discriminate a high and low tone by pressing the respective keys. All participants (N=30) completed two blocks without task order changes and thereafter six blocks in which task order randomly changed in 50% of the trials. The results revealed partial repetition costs for task order repetitions, but not for task order changes. Overall the results suggest that aspects of the abstract task-level structure and the element-level structure are stored as one conjoint memory episode.