This is the contextualized topic model trained on European Twitter data and used for cross-lingual topic analysis in the ICWSM paper "Bridging Nations: Quantifying the Role of Multilinguals in Communication on Social Media" (Mendelsohn et al., 2023).
Code for training and inferring topics: https://github.com/juliamendelsohn/bridging-nations/blob/main/src/topic_analysis/train_contextualized_topic_model.py More details about contextualized topic models: https://contextualized-topic-models.readthedocs.io/en/latest/introduction.html