This is a dataset repository which is used for model training, evalutaions as a part of our CTI research;

Threat Behavior Textual Search by Attention Graph Isomorphism (Bae et al., EACL 2024)

The dataset consists of pretraining dataset, threat reports per APT groups and a collector tool (which I use for all of this collection, needed to update new reports after our work).

• Textual corpus of threat reports

• Collected from 8 vendors

• A collection of threat reports by APT groups

• Our evaluation set is well-filtered, manually-verified set

• We also provide the copied list from two public websites (Malpedia, ThaiCERT).

• Our dataset is as of 2022. 06, we will be releasing our collector as a tool (working on... will be uploaded soon)

• Copyrights of all dataset belong to original authors or their vendors.

• Any misuse of attack information is strictly prohibited.

• Please contact us (Chanwoo Bae, [email protected]) for any questions.

• We kindly request to cite our paper with your use of dataset.