David Demitri Africa
Titles as the sources gave them. In 1 entry, with 0 checked quotations.
In the Base
Every entry they appear in.
Neologism Inoculation Prompting
According to the abstract, David Demitri Africa and Cameron Holmes propose neologism inoculation prompting, which swaps the plain-English inoculation prompt for a new trained token removed at deployment. It works on reward hacking but fails on sycophancy and harassment, and the authors say it hides behaviour rather than deleting it.