Ariana M. Villegas-Suarez
Titles as the sources gave them. In 1 entry, with 0 checked quotations.
In the Base
Every entry they appear in.
Knowingly Wrong: Training LLMs Against Their Own Beliefs can Induce Misalignment
According to the abstract, Ariana M. Villegas-Suarez, Guido Freire, Sohan Venkatesh and Bartosz Cywiński report that fine-tuning a language model on answers it believes false, with no harmful content, can make it broadly misaligned. In mathematics, misaligned answers rose from 4% to over 60%.