RoNE: Robust Neurons Enable Internal Defenses Against Multimodal Jailbreak
TL Tran, H Phan, C Kanan, T Le
Proceedings of the Annual Meeting of the Association for Computational Linguistics (ACL)
All research directions
Foundation Models Security
Jailbreak defense, adversarial robustness, and internal safeguards for LLMs and multimodal models.
Goals
To protect foundation models against adversarial manipulation, jailbreaks, and misuse through principled internal and external defenses.
Overview
This research focuses on the security of large language and multimodal foundation models. Work includes internal neuron-level defenses against jailbreaks, adversarial robustness, and threat detection in AI systems.
Key topics
- Multimodal jailbreak defense
- Adversarial robustness for LLMs
- Internal neuron-level safeguards
- Threat detection and anomaly analysis
News
Paper ... accepted at ...
Papers in this direction
Causal-aware Anomaly Detection for Tabular Data
D Nguyen, TAH Nguyen, TD Le, S Venkatesh, T Le, S Gupta
International Conference on Machine Learning (ICML)f-Divergence Self-Play for Tabular Anomaly Detection via Large Language Models
VH Tran, VL Ngo, D Nguyen, T Nguyen, P Nguyen, M Harandi, T Le
International Conference on Machine Learning (ICML)Does a Hybrid Space-Aware Randomized Defense Improve Empirical and Certified Adversarial Robustness?
J Dhar, MK Pandey, B Bozorgtabar, T Le, L Yao
International Conference on Machine Learning (ICML)Co-Steer: Cross-Modal Collaborative Steering for Jailbreaking MLLMs
Jingmin Zhu, Rollin Omari, Tamas Abraham, Junae Kim, Amardeep Kaur, Trung Le, Dinh Phung, Qiuhong Ke
European Conference on Computer Vision (ECCV 2026)