Here we share the full experimental data including the train/validation/test/sampled-negatives-for-evaluation/user-and-item-profiles for three sampled slices of Amazon and Goodreads datasets* used in our paper:
Ghazaleh Haratinezhad Torbati, Anna Tigunova, Gerhard Weikum. Unveiling Challenging Cases in Text-based Recommender Systems. In Proceedings of the Perspectives on the Evaluation of the Recommender Systems 3rd Workshop at RecSys 2023.
* The source of the original datasets are: https://nijianmo.github.io/amazon/index.html