Session Files used for the Sim4IA 2025 Micro-Shared-Task

This Zenodo record provides the training and test session datasets used in the Sim4IA Micro-Shared Tasks A and B, which focus on next query or utterance prediction. The sessions were extracted from log files provided by CORE. These datasets were specifically curated to support the development and evaluation of systems capable of anticipating the user's next input in interactive information-seeking scenarios. Task A targets the prediction of the next query in a search session, while Task B extends this to utterances within a broader conversational context. The data is intended to facilitate research on session-based and conversational information retrieval models.