Pseudo-relevance feedback mechanisms such as relevance models have shown the impact of expanding and re-weighting the users' initial queries exploiting information derived from in an initial set of retrieved documents, known as the pseudo-relevant documents. retrieval - through the use of pre-trained language models such as BERT to compute the relevance scores of queries and documents from their contents - has produced significant improvements in the efectiveness of several information retrieval tasks. Up so far, two diferent dense retrieval families have emerged: the use of single embedded representations for each passage and query, or via multiple representations, for each token in each passage and query. In this talk, we will discuss the first study into the potential for multiple representation dense retrieval to be enhanced using pseudo-relevance feedback, jointly carried out by the University of Pisa and the University of Glasgow.