Julian Michael
Research Scientist, Meta
Joined
July 2019
Names
Julian Michael
Emails
****@cs.washington.edu (Confirmed)
****@gmail.com (Confirmed)
****@nyu.edu (Confirmed)
****@nyu.edu (Confirmed)
****@scale.com (Confirmed)
****@meta.com (Confirmed)
Personal Links
Career & Education History
Research Scientist
Meta (meta.com)
2025 – Present
Head of SEAL
Scale AI (scale.com)
2025 – 2025
Postdoc
Center for Data Science, New York University (nyu.edu)
2022 – 2024
PhD student
Computer Science & Engineering, University of Washington (cs.washington.edu)
2015 – 2022
Undergrad student
Department of Computer Science, University of Texas at Austin (cs.utexas.edu)
2011 – 2015
Advisors, Relations & Conflicts
Expertise
alignment
, truthfulness
, debate
2022 – Present
interpretability
, explainability
2019 – Present
benchmarks
, evaluation
, linguistic analysis
2018 – Present
semantics
, semantic roles
, semantic formalisms
2015 – Present
structure prediction
, parsing
, ccg
2015 – Present
Publications
LLM Novice Uplift on Dual-Use, In Silico Biology Tasks: A Multi-Benchmark Assessment
Chen Bo Calvin Zhang, Christina Q Knight, Nicholas Kruus, Jason Hausenloy, Nathaniel Li, Aiden Kim, Yury Orlovskiy, Coleman Breen, Bryce Cai, Jasper Götting, Andrew Bo Liu, Samira Nedungadi, Paula Rodriguez, Yannis Yiming He, Zifan Wang, Seth Donoughe, Julian Michael- ICLR 2026 AIWILD
- Readers: Everyone
Best Practices for Biorisk Evaluations on Open-Weight Bio-Foundation Models
Boyi Wei, Zora Che, Nathaniel Li, Jasper Götting, Samira Nedungadi, Julian Michael, Summer Yue, Dan Hendrycks, Peter Henderson, Zifan Wang, Seth Donoughe, Mantas Mazeika- BioSafe GenAI 2025 Poster
- Readers: Everyone
Search-Time Data Contamination
- NeurIPS 2025 LLM Evaluation Workshop Poster
- Readers: Everyone
Inverse Scaling in Test-Time Compute
Aryo Pradipta Gema, Alexander Hägele, Runjin Chen, Andy Arditi, Jacob Goldman-Wetzler, Kit Fraser-Taliente, Henry Sleight, Linda Petrini, Julian Michael, Beatrice Alex, Pasquale Minervini, Yanda Chen, Joe Benton, Ethan Perez- Accepted by TMLR
- Readers: Everyone
Teaching Models to Verbalize Reward Hacking in Chain-of-Thought Reasoning
- ICML 2025 R2-FM Workshop Poster
- Readers: Everyone
Quantifying Elicitation of Latent Capabilities in Language Models
Elizabeth Donoway, Hailey Joren, Arushi Somani, Henry Sleight, Julian Michael, Michael R DeWeese, John Schulman, Ethan Perez, Fabien Roger, Jan Leike- NeurIPS 2025 poster
- Readers: Everyone
Why Do Some Language Models Fake Alignment While Others Don't?
- NeurIPS 2025 spotlight
- Readers: Everyone
AI Debate Aids Assessment of Controversial Claims
Salman Rahman, Sheriff Issaka, Ashima Suvarna, Genglin Liu, James Shiffer, Jaeyoung Lee, Md Rizwan Parvez, Hamid Palangi, Shi Feng, Nanyun Peng, Yejin Choi, Julian Michael, Liwei Jiang, Saadia Gabriel- NeurIPS 2025 poster
- Readers: Everyone
Bias-Augmented Consistency Training Reduces Biased Reasoning in Chain-of-Thought
- ICLR 2025 Conference Withdrawn Submission
- Readers: Everyone
Evaluating Oversight Robustness with Incentivized Reward Hacking
- ICLR 2025 Conference Withdrawn Submission
- Readers: Everyone
Co-Authors
- Aaron Mueller
- Abhay Sheshadri
- Aiden Kim
- Akbir Khan
- Alane Suhr
- Alex Infanger
- Alex Troy Mallen
- Alex Wang
- Alexander Hägele
- Alexander Koller
- Alicia Parrish
- Alisa Liu
- Alwin Peng
- Amanpreet Singh
- Amelia Harrison
- Amelia J. Harrison
- Andrew Bo Liu
- Andy Arditi
- Angelica Chen
- Ari Holtzman
- Arun Jose
- Arushi Somani
- Aryo Pradipta Gema
- Asa Cooper Stickland
- Ashima Suvarna
