This paper addresses the problem of discovering conversa- tional group dynamics from nonverbal cues extracted from thin-slices of interaction. We first propose and analyze a novel thin-slice interaction descriptor - a bag of group non- verbal patterns - which robustly captures the turn-taking behavior of the members of a group while integrating its leader’s position. We then rely on probabilistic topic mod- eling of the interaction descriptors which, in a fully unsu- pervised way, is able to discover group interaction patterns that resemble prototypical leadership styles proposed in so- cial psychology. Our method, validated on the Augmented Multi-Party Interaction (AMI) meeting corpus, facilitates the retrieval of group conversational segments where seman- tically meaningful group behaviours emerge, without the need of any previous labeling.