The go/no‐go with compound stimuli is an alternative to matching‐to‐sample to produce conditional and emergent relations in adults. The aim of this study was to evaluate the effectiveness of this procedure with two children diagnosed with autism. We trained and tested participants to respond to conditional relations among arbitrary stimuli using the go/no‐go procedure. Both learned all the trained conditional relations without developing response bias or responding to no‐go trials. Participants demonstrated performance consistent with symmetry, but not equivalence.