Cognitive Bias Modification–Interpretation (CBM-I) training has been put forward as a promising new intervention for youth with psychopathology. A recent Randomised Controlled Trial (RCT) showed that an online CBM-I training designed to reduce dysfunctional interpretations in youth with Obsessive Compulsive Disorder (OCD) had therapeutic benefits on OCD symptoms. In addition, there are practical benefits as the online and automated nature of the training allows for 24/7 accessibility, is cheap and an easy to implement intervention. There is, however, significant variability in CBM-I training effects on symptoms. By conducting secondary analyses of the online CBM-I RCT, we aimed to examine whether baseline OCD severity, interpretation bias, and degree of autism symptoms are related to training effectiveness. In the RCT, 36 children with OCD (8–18 years) followed 12-sessions CBM-I training. Bayesian analyses showed no evidence for any of the three predictors being associated with CBM-I effects on OCD symptoms. These results offer no answer to the question for whom CBM-I training works best. However, there is also no evidence that CBM-I might work less well for these subgroups. Future research with larger samples is necessary to test the robustness of these findings.