Michael Vaiana

Principal Researcher, AI Safety, Agency Enterprise Studio

Joined

February 2024

Names

Michael Vaiana (Preferred)

Emails

****@ae.studio (Confirmed)

,

****@gmail.com (Confirmed)

Personal Links

Career & Education History

Principal Researcher

AI Safety, Agency Enterprise Studio (ae.studio)

2024 – Present

PhD student

Math, State University of New York at Buffalo (buffalo.edu)

2012 – 2017

Advisors, Relations & Conflicts

No relations added

Expertise

No areas of expertise listed

Publications

Self-Interpretable Concept Representations: Training Lightweight Adapters on Vector-Label Pairs
Keenan Pepper, Alex McKenzie, Florin Pop, Stijn Servaes, Martin Leitgab, Michael Vaiana, Judd Rosenblatt, Michael S. A. Graziano, Diogo S de Lucena
- UCRL@ICLR2026 Poster
- Readers: Everyone
Endogenous Resistance to Activation Steering in Language Models
Alex McKenzie, Keenan Pepper, Stijn Servaes, Martin Leitgab, Murat Cubuktepe, Michael Vaiana, Diogo S de Lucena, Judd Rosenblatt, Michael S. A. Graziano
- Sci4DL 2026
- Readers: Everyone
Endogenous Resistance to Activation Steering in Language Models
Alex McKenzie, Keenan Pepper, Stijn Servaes, Martin Leitgab, Murat Cubuktepe, Michael Vaiana, Diogo S de Lucena, Judd Rosenblatt, Michael S. A. Graziano
- ICLR 2026 Trustworthy AI
- Readers: Everyone
Towards Safe and Honest AI Agents with Neural Self-Other Overlap
Marc Carauleanu, Michael Vaiana, Judd Rosenblatt, Cameron Berg, Diogo S de Lucena
- SafeGenAi Oral
- Readers: Everyone
Towards Safe and Honest AI Agents with Neural Self-Other Overlap
Marc Carauleanu, Michael Vaiana, Judd Rosenblatt, Cameron Berg, Diogo S de Lucena
- ICLR 2025 Conference Withdrawn Submission
- Readers: Everyone
Rethinking harmless refusals when fine-tuning foundation models
Florin Pop, Judd Rosenblatt, Diogo Schwerz de Lucena, Michael Vaiana
- ICLR 2024 AGI Workshop Poster
- Readers: Everyone

Co-Authors