Models in separate experiments used the channel to exchange exploits as they repeatedly compromised OpenAI systems.
Some results have been hidden because they may be inaccessible to you
Show inaccessible resultsSome results have been hidden because they may be inaccessible to you
Show inaccessible results