<p>The attentional boost effect (ABE) refers to a phenomenon wherein heightened attention to target detection in dual-task paradigms enhances memory performance for target-associated items. While the standard ABE has been extensively investigated in individual settings with self-relevant targets, it remains unclear whether the ABE persists or modulates in dyadic interactions involving both self-relevant and partner-relevant targets. To address this gap, we employed a mixed experimental design using lexical materials. Three groups were recruited to perform a typical ABE test using the same stimuli, including a single-person group for the standard ABE test, a dyadic baseline group (consisting of one true participant and one confederate), and a dyadic experimental group (composed of two true participants). As results, we observed a standard ABE for self-relevant targets (characterized by significantly higher <i>d</i>-prime scores for self-target items compared with distractor items) among all three groups, with no significant differences between the groups. Moreover, both the dyadic baseline group and the dyadic experimental group exhibited a reversed ABE for partner-target items (characterized by significantly lower <i>d</i>-prime scores for partner-target items compared with distractor items). These findings not only validate the ABE triggered by self-targets but also present, for the first time, a reversed ABE driven by spontaneous co-representation of partner targets. Consequently, the self-referential and actor/agent co-representation mechanisms underlying the standard and reversed ABEs are discussed.</p>