David Africa

Karma: 269

Research Scientist with the Alignment team at UK AISI.

Inoculation prompting: Instructing models to misbehave at train-time can improve run-time behavior

Sam Marks, Nevan Wichers, Daniel Tan, Aram Ebtekar, Jozdien, David Africa, Alex Mallen and Fabien Roger

8 Oct 2025 22:02 UTC

152 points

37 comments2 min readLW link

Subliminal Learning, the Lottery-Ticket Hypothesis, and Mode Connectivity

David Africa6 Oct 2025 15:26 UTC

23 points

6 comments7 min readLW link

No Answer Needed: Predicting LLM Answer Accuracy from Question-Only Linear Probes

antonghawthorne, ivanvmoreno, Arnau Padrés Masdemont, David Africa and LorenzoPacchiardi

16 Sep 2025 15:23 UTC

9 points

0 comments4 min readLW link

(arxiv.org)

Large Language Models and the Critical Brain Hypothesis

David Africa9 Sep 2025 15:45 UTC

33 points

0 comments6 min readLW link

Research Areas in Learning Theory (The Alignment Project by UK AISI)

David Africa and Edmund Lau

1 Aug 2025 10:26 UTC

15 points

0 comments24 min readLW link

(alignmentproject.aisi.gov.uk)

The Alignment Project by UK AISI

Mojmir, Benjamin Hilton, Jacob Pfau, Geoffrey Irving, Joseph Bloom, Tomek Korbak, David Africa and Edmund Lau

1 Aug 2025 9:52 UTC

29 points

0 comments2 min readLW link

(alignmentproject.aisi.gov.uk)