from Hacker News

Come down a rabbit hole with OpenAI o1 system card

by roh26it on 12/16/24, 12:18 PM with 1 comments

  • by roh26it on 12/16/24, 12:18 PM

    The most interesting findings (at least from my perspective) - OpenAI measured "honesty" that could become an interesting model eval others should adopt. - Using instruction heirarchy to _avoid_ prompt injections - When pushed to pursue misaligned goals, it tried to disable oversight in 5% of tests!