2

Is GPT3 adversarial? And if not, has there been any work to combine GPT3 with GANs, specifically to improve plausibility or "humanness" or otherwise tune the personality or style of response?

Comments
  • 1
    I can't remember the architecture behind it described in blogposts at the time, but IIRC, it is indeed adversarial (at least in parts).
Add Comment