Four experienced group therapists and 50 group counseling trainees indicated which of the 19 group counseling interventions they would consider using to address 21 different group counseling situations depicted in the Group Therapy Questionnaire (GTQ; Wile, Brown, & Pollack, 1970). Pathfinder Network Analyses (Schvaneveldt, 1990) was used to create knowledge structures (cognitive maps) for group leader interventions from participants' responses to the GTQ. These knowledge structures consisted of 19 nodes, representing the group counseling interventions, and links representing the relationships among these interventions. We identified errors of commission (links in trainees' knowledge structures not found experts' knowledge structures) and errors of omission (links in experts' knowledge structures not present in trainees' knowledge structures). At least 75% of the trainees had nine frequently occurring errors of omission, but only one frequently occurring error of commission (silence to attack). Cluster analysis identified four subsets of trainees that had the same frequent errors of omission and commission. Based on the specific errors present, the subsets of trainees were labeled: Expansive/Confused about Attacking Learners, Undifferentiated but Discerning Learners, Narrow/Passive Learners, and Wide Ranging/Incorrect about Group Processes Learners. We argue that knowing the type of errors in trainees' implicit knowledge structures can help group counseling trainers better target their instructional activities.