All three datasets were built on the basis of independent snapshots of the BEAR datasets. Each triple was together with a creation and deletion timestamp cast into a nested quoted triples in subject position. The timestamps for both datasets do not reflect the original timestamps of the triples but were for the sake of evaluating the RDF-star policies incremented by 1sec from the initial system timestamp. The main script used to construct these datasets can be found here, which contains the logic on how to assemble these datasets. However, the script was run within a Docker container found on the starvers_eval Github page.