<p>In addition to its well-known association with mental effort, pupil dilation provides insight into the cognitive and neurobiological basis of episodic memory. Notably, pupil responses during retrieval are influenced by the encoding conditions under which the information was originally learned. In our study, we manipulated encoding by presenting participants with word pairs either once or twice, and then examined pupil responses during subsequent retrieval tasks that employed either a recognition or a cued recall paradigm. Repeated encoding led to higher accuracy and faster response times, confirming enhanced memory performance across both tasks. Critically, however, the strengthening of memory traces was accompanied by reduced pupil dilation only in the cued recall task, suggesting that stronger memories required less cognitive effort during the demanding recall process. In contrast, during the recognition task, despite a robust pupil old/new effect (e.g. increased pupil responses to correct “old” responses relative to correct “new” responses) the magnitude of pupil dilation was not modulated by memory trace strength. This implies that, in recognition decisions, pupil responses may primarily reflect the recollection–familiarity dichotomy rather than memory trace strength. Overall, our findings demonstrate that the type of memory test influences how effectively pupillometry can assess memory trace strength.</p>