Cognitive control processes mirror fast and dynamic adaptation toward a change in the environment. When performing dual tasks, mental representations of dual-task-specific control requirements and the task-pair set are established that help to manage dual-task processing (Hirsch et al., 2017, 2018; Hommel, 2004, 2020). In the present study, we investigated to which extent such higher order representations of dual-task processing persist even if major characteristics of the task context change, for example, if one of the tasks of a dual task becomes irrelevant. For this, we adapted the fade-out paradigm (Mayr & Liebscher, 2001) to a dual-task setting and tested whether fade-out costs appear. Performance of pure Task 1 single tasking was compared to the performance of Task 1 processing right after dual-task trials (fade-out phase). Results showed that performance in this fade-out block did not immediately drop to single-task performance (fade-out costs), indicating the persistence of task-pair set representations (Experiments 1 and 3, N = 40 each). In addition, automatic stimulus-response translation processes continued within the fade-out phase, resulting in ongoing between-task interference. Furthermore, the frequency of between-task interference in dual-task blocks was manipulated (75% vs. 25% incongruence) between participants to establish conflict-biased control states of increased versus relaxed task shielding. These different control states, however, did not modulate fade-out costs (Experiment 2, N = 80). Nevertheless, the persistence of these control adaptations was reflected in manipulation-dependent between-task interference during fade-out trials. Implications of this new evidence are discussed. (PsycInfo Database Record (c) 2024 APA, all rights reserved).