The persistent belief in misinformation despite correction, known as the Continued Influence Effect (CIE), is a major challenge for interventions that aim to mitigate risks from harmful content. Social media plays a crucial role in spreading misinformation, but also in its correction. A promising approach to mitigate the influence of misinformation are community notes, a crowd-sourced fact-checking tool implemented by X. Community notes provide corrective context directly with the tweet containing the misinformation. This study investigates whether community notes are perceived as correction and whether they could reduce the continued use of misinformation across neutral and socially relevant content. To this end, N = 542 participants from a quota-based German sample (age, gender) read short narratives presented as tweets with neutral (e.g., food recall) or socially relevant content (e.g. gender quota) in a realistic social media environment. Within each narrative one tweet contained the misinformation and was later presented a second time alongside a community note. The community note contained either a simple retraction of the misinformation or the correction with an alternative explanation. For neutral content, we replicated the typical CIE pattern where a retraction including an alternative explanation was most effective in reducing CIE, followed by a simple retraction. For socially relevant content, community notes only reduced the continued use of misinformation when providing a retraction with an alternative explanation. These findings suggest that community notes can reduce the influence of misinformation on social media, although their effectiveness may be moderated by ideological factors in socially relevant contexts.