Lucius Bushnaq
Bio
Lucius Bushnaq is an AI safety researcher specializing in mechanistic interpretability, currently a researcher at Goodfire AI. He holds a PhD in mathematics from Trinity College Dublin (2020-2023) and an MS in physics from RWTH Aachen. He co-founded Apollo Research, where he led interpretability work including the Local Interaction Basis method and degeneracy-based approaches to mechanistic interpretability. His research trajectory began with work on selection theorems for modularity in neural networks, conducted through AI Safety Camp under the mentorship of John Wentworth, which was supported by Long-Term Future Fund grants. He has published papers on parameter decomposition, LLM modularity, and identifying computationally relevant sparse features in neural networks.
Grants
from Long-Term Future Fund
from Long-Term Future Fund
Discussion
Sign in to join the discussion.
No comments yet. Be the first to share your thoughts.
Details
- Last Updated
- Mar 22, 2026, 11:03 PM UTC
- Created
- Mar 20, 2026, 2:54 AM UTC