The value of initial condition ensembles for characterizing uncertainty related to internal variability of the climate system has been well established. E3SMv1 and E3SMv2 have made use of such large ensemble experiments and E3SMv3 will continue that effort. E3SMv3 will build upon this initial condition ensemble design by adding a complementary dimension of uncertainty – climate sensitivity – which is expected to be a critical component of quantifying future changes. Making use of an AI-based calibration effort, alternative parameter sets are being identified which alter the climate sensitivity in E3SMv3, while maintaining a credible representation of the present-day climate. In this work we will provide preliminary results from the candidate selection process.