Recent analyses of human creativity and curiosity have identified the existence of three styles of exploration: busybody, hunter, and dancer. These styles were recognized largely by observing participants’ explorations within a task, converting those observations into networks, and measuring networks’ properties. But do these exploration styles still appear across different tasks? And when graph-based descriptors of an individual’s exploration style are identified, how well do they transfer to similar tasks? We study inter- and intra-individual differences in two similar, but distinct, word association tasks: Chain Free Association and Semantic Fluency. We demonstrate that in some cases, graph-theoretic features do seem to capture individual semantic exploration patterns across tasks. Furthermore, our results provide evidence supporting the existence of the dancer style and its relationship to the Busybody-Hunter score. These findings highlight the potential of graph analysis as a tool for characterizing and exploring individual cognitive styles in semantic tasks.