Rex ModernBERT Investigation Engine (RBA Pre-trained)

This checkpoint is a specialized ModernBERT Dual-Head MECE Investigation Engine trained on the Risk-Based Authentication (RBA) dataset with Popperian Process of Elimination (PoE).

Model Architecture:

• Base Encoder: answerdotai/ModernBERT-base (Apache-2.0)
• Heads: ToolMLPHead (4 Investigation Slots) & SharedVerdictHead (4 MECE Root Causes)
• Hypotheses: H_botnet, H_sybil, H_shared, H_novel
• Investigation Policy: 3-Round Dynamic Budgeted Telemetry Collection

Training Dataset & Citation:

This model was trained using data derived from the Login Data Set for Risk-Based Authentication hosted on Zenodo (Record 6782156), licensed under Creative Commons Attribution 4.0 International (CC BY 4.0):

Stephan Wiefling, Paul René Jørgensen, Sigurd Thunem, and Luigi Lo Iacono. 2022. Pump Up Password Security! Evaluating and Enhancing Risk-Based Authentication on a Real-World Large-Scale Online Service. In ACM Transactions on Privacy and Security (TOPS), Vol. 26, No. 1, Article 7. DOI: https://doi.org/10.1145/3546187. Dataset: https://doi.org/10.5281/zenodo.6782156