Rohin Shah

Researcher, Google DeepMind

Joined

April 2018

Names

Rohin Shah

Emails

****@berkeley.edu (Confirmed)

****@gmail.com (Confirmed)

****@deepmind.com (Confirmed)

****@google.com (Confirmed)

Personal Links

Homepage

Google Scholar

DBLP

Career & Education History

Researcher

Google DeepMind (deepmind.com)

2020 – Present

PhD student

University of California Berkeley (berkeley.edu)

2014 – 2020

Advisors, Relations & Conflicts

PhD Advisor

Stuart Russell

2017 – 2020

PhD Advisor

Anca Dragan

2017 – 2020

PhD Advisor

Pieter Abbeel

2017 – 2020

Expertise

ai safety

Present

preference learning

Present

ai alignment

Present

interpretability

Present

Publications

A Pragmatic Way to Measure Chain-of-Thought Monitorability
Roland S. Zimmermann, Scott Emmons, Rohin Shah, David Elson
- OpenReview Archive Direct Upload
- Readers: Everyone
Evaluating LLM Goal-Directedness
Tom Everitt, Cristina Garbacea, Alexis Bellot, Jonathan Richens, Henry Papadatos, Siméon Campos, Rohin Shah
- Submitted to ICLR 2026
- Readers: Everyone
Consistency Training Helps Stop Sycophancy and Jailbreaks
Alex Irpan, Alexander Matt Turner, Mark Kurzeja, David Elson, Rohin Shah
- Submitted to ICLR 2026
- Readers: Everyone
Evaluating Frontier Models for Stealth and Situational Awareness
Mary Phuong, Roland S. Zimmermann, Ziyue Wang, David Lindner, Victoria Krakovna, Sarah Cogan, Allan Dafoe, Lewis Ho, Rohin Shah
- Submitted to NeurIPS 2025
- Readers: Everyone
MONA: Myopic Optimization with Non-myopic Approval Can Mitigate Multi-step Reward Hacking
Sebastian Farquhar, Vikrant Varma, David Lindner, David Elson, Caleb Biddulph, Ian Goodfellow, Rohin Shah
- ICML 2025 poster
- Readers: Everyone
MONA: Myopic Optimization with Non-myopic Approval Can Mitigate Multi-step Reward Hacking
Sebastian Farquhar, Vikrant Varma, David Lindner, David Elson, Caleb Biddulph, Ian Goodfellow, Rohin Shah
- CoRR 2025
- Readers: Everyone
Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2
Tom Lieberum, Senthooran Rajamanoharan, Arthur Conmy, Lewis Smith, Nicolas Sonnerat, Vikrant Varma, Janos Kramar, Anca Dragan, Rohin Shah, Neel Nanda
- BlackboxNLP 2024
- Readers: Everyone
BEDD: The MineRL BASALT Evaluation and Demonstrations Dataset for Training and Benchmarking Agents that Solve Fuzzy Tasks
Stephanie Milani, Anssi Kanervisto, Karolis Jucys, Sander V Schulhoff, Brandon Houghton, Rohin Shah
- MFM-EAI@ICML2024 Poster
- Readers: Everyone
Improving Sparse Decomposition of Language Model Activations with Gated Sparse Autoencoders
Senthooran Rajamanoharan, Arthur Conmy, Lewis Smith, Tom Lieberum, Vikrant Varma, Janos Kramar, Rohin Shah, Neel Nanda
- ICML 2024 MI Workshop Poster
- Readers: Everyone
Improving Sparse Decomposition of Language Model Activations with Gated Sparse Autoencoders
Senthooran Rajamanoharan, Arthur Conmy, Lewis Smith, Tom Lieberum, Vikrant Varma, Janos Kramar, Rohin Shah, Neel Nanda
- NeurIPS 2024 poster
- Readers: Everyone

View all 62 publications

Co-Authors

View all 154 co-authors