RussellThor comments on Should we publish mechanistic interpretability research?

RussellThor 22 Apr 2023 5:49 UTC
1 point
0
“Most people who work on ML do not care about alignment” I am a bit surprised by this, if it was true, are you sure it still is?