TRITIUM

Malwarebazaar dataset from ~2022 threat actors (used in Rapidrift study)

The dataset is a .h5 file containing both the feature vector and metadata as a single Dataframe which can be loaded using pandas (key='xy').

tritium.h5 contains

~23k samples (inclusive of malicious and benign)

tritium_unseen.h5 contains

~14k samples from malware families not seen in BODMAS

Link to dataset(s): https://drive.google.com/drive/folders/1KGeUYS7SKCJprYJQoqPqew0DXnaDGQgz?usp=sharing

Request access to the Rapidrift framework and the original malware samples

If you are interested in using the framework demonstrated in the study and/or would like to access the original malware samples for this dataset; kindly drop a message to [email protected] using your work/academic institution email.