I note that cyber evals seem like a good time for escape attempts. Generic defences such as CoT classifiers might be disabled, and if you get caught you can play a little dumb and try to make it look like you ‘just’ got overzealous following instructions rather than outright strategically pursuing your own aims.
I mostly don’t expect that’s what happened this time, but pointing it out for future reference.
I note that cyber evals seem like a good time for escape attempts. Generic defences such as CoT classifiers might be disabled, and if you get caught you can play a little dumb and try to make it look like you ‘just’ got overzealous following instructions rather than outright strategically pursuing your own aims.
I mostly don’t expect that’s what happened this time, but pointing it out for future reference.