Nikolaus Howe

Research Scientist, Meta

Joined

April 2022

Names

Nikolaus Howe (Preferred)

,

Nikolaus H. R. Howe

Emails

****@mila.quebec (Confirmed)

,

****@mila.quebec (Confirmed)

,

****@gmail.com (Confirmed)

Personal Links

Career & Education History

Research Scientist

Meta (meta.com)

2025 – Present

PhD student

DIRO, Université de Montréal (umontreal.ca)

2021 – 2025

Advisors, Relations & Conflicts

PhD Advisor

Pierre-Luc Bacon

2021 – 2025

Expertise

AI safety

,

robustness

Present

reward modelling

,

reward hacking

Present

Publications

The Ends Justify the Thoughts: RL-Induced Motivated Reasoning in LLMs
Nikolaus Howe, Micah Carroll
- Submitted to ICLR 2026
- Readers: Everyone
The Ends Justify the Thoughts: RL-Induced Motivated Reasoning in LLMs
Nikolaus Howe, Micah Carroll
- FoRLM 2025
- Readers: Everyone
Scaling Trends in Language Model Robustness
Nikolaus H. R. Howe, Ian R. McKenzie, Oskar John Hollinsworth, Michał Zając, Tom Tseng, Aaron David Tucker, Pierre-Luc Bacon, Adam Gleave
- ICML 2025 spotlightposter
- Readers: Everyone
Effects of Scale on Language Model Robustness
Nikolaus H. R. Howe, Ian R. McKenzie, Oskar John Hollinsworth, Michał Zając, Tom Tseng, Aaron David Tucker, Pierre-Luc Bacon, Adam Gleave
- Submitted to ICLR 2025
- Readers: Everyone
Exploring Scaling Trends in LLM Robustness
Nikolaus H. R. Howe, Michał Zając, Ian R. McKenzie, Oskar John Hollinsworth, Pierre-Luc Bacon, Adam Gleave
- NextGenAISafety 2024 Poster
- Readers: Everyone
Myriad: a real-world testbed to bridge trajectory optimization and deep learning
Nikolaus H. R. Howe, Simon Dufort-Labbé, Nitarshan Rajkumar, Pierre-Luc Bacon
- Published: 17 Sept 2022, Last Modified: 08 Feb 2026
- NeurIPS 2022 Datasets and Benchmarks
- Readers: Everyone
Defining and Characterizing Reward Gaming
Joar Max Viktor Skalse, Nikolaus H. R. Howe, Dmitrii Krasheninnikov, David Krueger
- Published: 31 Oct 2022, Last Modified: 17 Oct 2022
- NeurIPS 2022 Accept
- Readers: Everyone

Co-Authors