Responses are generally faster for items that repeatedly require the same, previously-executed as opposed to a different response (repetition priming effect). This finding is explained by stimulus-response (S-R) associations which are formed when stimuli and responses co-occur and thus bind together. S-R associations can be divided into two components: Stimulus–action (S-A) associations between stimuli and motor outputs and stimulus–classification (S-C) associations between stimuli and their task-specific semantic classification. In the present work, we tested the impact of instructing participants to intentionally remember or forget the stimulus at encoding and/or retrieval of associated actions and classifications. We discuss five attempts to obtain directed forgetting of repetition priming effects using the list-method method (Experiments 1 + 2) as well as the item-method of directed forgetting (Experiments 3-5). The absence of S-A repetition priming effects in half of our experiments suggests that an additional memory instruction hinders the formation of robust S-A associations frequently observed in other studies. This finding implies crosstalk between declarative (remember the stimulus) and procedural (act upon the stimulus) memory. Further, despite variations in samples (lab + online experiments) and design, we observed no standard directed forgetting effects for incidentally-learned, stimulus-based associations. Thus, intending to forget a stimulus does not impair the retrieval or encoding of an associated action or classification. Rather, an additional memory task (remembering or forgetting) per se appears to interfere with the encoding of S-R associations.