Joseph Isaac Bloom

Researcher, UK AI Safety Institute, Department of Science, Technology and Innovation

Joined

May 2024

Names

Joseph Isaac Bloom (Preferred)

Emails

****@gmail.com (Confirmed)

****@student.unimelb.edu.au

****@massdynamics.com

Personal Links

Homepage

ORCID

Career & Education History

Researcher

UK AI Safety Institute, Department of Science, Technology and Innovation (dsit.gov.uk)

2024 – Present

Undergrad student

Mathematics, University of Melbourne (unimelb.edu.au)

2016 – 2019

Advisors, Relations & Conflicts

Coauthor

Neel Nanda

2023 – 2024

Expertise

Mechanistic Interpretability

2023 – Present

Sparse Autoencoders

2023 – Present

Publications

Auditing Games for Sandbagging
Jordan Taylor, Sid Black, Dillon Bowen, Thomas Read, Satvik Golechha, Alex Zelenka Martin, Oliver Makins, Connor Kissane, Kola Ayonrinde, Jacob Merizian, Samuel Marks, Chris Cundy, Joseph Isaac Bloom
- Arxiv
- Readers: Everyone
ContextBench: Modifying Contexts for Targeted Latent Activation and Behaviour Elicitation
Robert Graham, Edward Stevinson, Leo Richter, Alexander Chia, Joseph Miller, Joseph Isaac Bloom
- ICLR 2026 Poster
- Readers: Everyone
ContextBench: Modifying Contexts for Targeted Latent Activation and Behaviour Elicitation
Robert Graham, Edward Stevinson, Leo Richter, Alexander Chia, Joseph Miller, Joseph Isaac Bloom
- Mech Interp Workshop (NeurIPS 2025) Poster
- Readers: Everyone
A is for Absorption: Studying Feature Splitting and Absorption in Sparse Autoencoders
David Chanin, James Wilken-Smith, Tomáš Dulka, Hardik Bhatnagar, Satvik Golechha, Joseph Isaac Bloom
- NeurIPS 2025 oral
- Readers: Everyone
Open Problems in Mechanistic Interpretability
Lee Sharkey, Bilal Chughtai, Joshua Batson, Jack Lindsey, Jeffrey Wu, Lucius Bushnaq, Nicholas Goldowsky-Dill, Stefan Heimersheim, Alejandro Ortega, Joseph Isaac Bloom, Stella Biderman, Adrià Garriga-Alonso, Arthur Conmy, Neel Nanda, Jessica Mary Rumbelow, Martin Wattenberg, Nandi Schoots, Joseph Miller, William Saunders, Eric J Michaud et al. (9 additional authors not shown)
- Accepted by TMLR
- Readers: Everyone
SAEBench: A Comprehensive Benchmark for Sparse Autoencoders in Language Model Interpretability
Adam Karvonen, Can Rager, Johnny Lin, Curt Tigges, Joseph Isaac Bloom, David Chanin, Yeu-Tong Lau, Eoin Farrell, Callum Stuart McDougall, Kola Ayonrinde, Demian Till, Matthew Wearden, Arthur Conmy, Samuel Marks, Neel Nanda
- ICML 2025 poster
- Readers: Everyone
Interpreting Attention Layer Outputs with Sparse Autoencoders
Connor Kissane, Robert Krzyzanowski, Joseph Isaac Bloom, Arthur Conmy, Neel Nanda
- Submitted to ICLR 2025
- Readers: Everyone
A is for Absorption: Studying Feature Splitting and Absorption in Sparse Autoencoders
David Chanin, James Wilken-Smith, Tomáš Dulka, Hardik Bhatnagar, Joseph Isaac Bloom
- Submitted to ICLR 2025
- Readers: Everyone
Sparse Autoencoders Do Not Find Canonical Units of Analysis
Patrick Leask, Bart Bussmann, Michael T Pearce, Joseph Isaac Bloom, Curt Tigges, Noura Al Moubayed, Lee Sharkey, Neel Nanda
- ICLR 2025 Poster
- Readers: Everyone
Stitching Sparse Autoencoders of Different Sizes
Patrick Leask, Bart Bussmann, Joseph Isaac Bloom, Curt Tigges, Noura Al Moubayed, Neel Nanda
- SciForDL Poster
- Readers: Everyone

View all 12 publications

Co-Authors

View all 62 co-authors