Scholar
David Demitri Africa
Google Scholar ID: JsydmjkAAAAJ
UK AI Security Institute (AISI)
Alignment
Follow
Homepage
↗
Google Scholar
↗
Citations & Impact
All-time
Citations
8
H-index
2
i10-index
0
Publications
10
Co-authors
12
list available
Contact
Email
David.Africa@dsit.gov.uk
Twitter
Open ↗
GitHub
Open ↗
Publications
14 items
Prefill Awareness in Large Language Models
2026
Cited
0
Consistency Training Along the Transformer Stack
2026
Cited
0
Consistency Training Can Entrench Misalignment
2026
Cited
0
Consistency Training while Mitigating Obfuscation via Rate Matching
2026
Cited
0
Does Self-Evaluation Enable Wireheading in Language Models?
2025
Cited
0
Inoculation Prompting: Eliciting traits from LLMs during training can suppress them at test-time
2025
Cited
0
Pico: A Modular Framework for Hypothesis-Driven Small Language Model Research
2025
Cited
0
Investigating ReLoRA: Effects on the Learning Dynamics of Small Language Models
2025
Cited
0
Load more
Resume (English only)
Research Experience
Research Scientist with the Alignment team at the UK AI Security Institute. Jardine Postgraduate Cambridge Scholar.
Background
Interested in technical AI safety, natural language processing, and collective behavior.
Miscellany
Used to do competitive debate. Trying to get good at climbing. Next planned hike: Mt. Kabunian.
Co-authors
12 total
Richard Diehl Martinez
PhD Student, University of Cambridge
Suchir Salhan
University of Cambridge
Co-author 3
JIMSON PAULO LAYACAN
Ateneo de Manila University
Lorenzo Pacchiardi
Research Associate, University of Cambridge
Ryan Daniels
University of Cambridge
Niels Warncke
Center on Longterm-Risk
Daniel Tan
UCL
×
Welcome back
Sign in to Agora
Welcome back! Please sign in to continue.
Email address
Password
Forgot password?
Continue
Do not have an account?
Sign up