Scholar
Bilal Chughtai
Google Scholar ID: i-L98bwAAAAJ
Google DeepMind
Citations & Impact
Citations
355
H-index
8
i10-index
7
Publications
10
Co-authors
24
list available
Publications
Resume (English only)
Academic Achievements
- Book Summary: Zero to One
- Detecting strategic deception using linear probes
- Open problems in mechanistic interpretability
- Activation space interpretability may be doomed
- Understanding positional features in layer 0 SAEs
- Unlearning via RMU is mostly shallow
- Transformer circuit faithfulness metrics are not robust
- Me, Myself, and AI: The Situational Awareness Dataset (SAD) for LLMs
Research Experience
- Research Engineer @ Google DeepMind
Background
- Research interests include language model interpretability and AGI safety. Aiming to make the development of transformative AI go well for humanity.
Miscellany
- Personal interests include writing and sharing other interests on this blog.
