Memory is composed of three stages: acquisition, consolidation, and retrieval. By impairing acquisition processes, benzodiazepines cause anterograde amnesia while leaving intact information learned before the drug was taken. In some circumstances, retrieval of this information is even improved by benzodiazepines. It has been hypothesized that this phenomenon is not a true facilitation of retrieval processes, but is the result of reduced interference from items presented after drug administration and is thus a secondary consequence of drug-induced amnesia. Experiment 1 investigated the effect of 0.5, 1, and 2.5 mg of lorazepam on explicit episodic memory in healthy young volunteers. The 1-mg dose was found to significantly improve recall of items presented before drug administration without causing amnesia for items presented after drug administration, thus excluding an interference explanation. Experiment 2 investigated the conditions necessary to obtain facilitated retrieval with 1 mg of lorazepam. The results showed that facilitation was found only when two lists of semantically related material were presented, but that both of the lists could be presented before drug administration, thus excluding an effect of lorazepam on consolidation. Facilitation could be demonstrated in both direct (free recall) and indirect (backwards reading) retrieval tasks and when all of the material was presented after lorazepam administration. This improved retrieval could therefore be of clinical relevance, but any benefits would be reduced at higher doses that at the same time impair acquisition of new information. However, because 1 mg of lorazepam is an effective anxiolytic dose, these results suggest that it is possible to combine effective anxiety reduction with some benefits to memory.