4 ms·

They finished training GPT4 mid last year, before ChatGPT was released

• z3c0 4y ago
  There's definitely some "Mechanical Turk"-ing going on behind the scenes. I enjoy the ChatGPT jailbreaking scene, and almost every time there's a significant break found, within a day, ChatGPT will dance around it (take the big token-leak jailbreak from a few days ago.) The GPT4 model may not be retrained itself, but it seems fairly evident that the ability for supplemental rules/fine-tuning is baked into their architecture for ChatGPT.
  • janalsncm 4y ago
    Even with mechanical Turking it’s still an impressive claim. I have not seen any experiments which demonstrate that autoregressive language models can reliably play at any level, let alone superhuman levels. Until I see evidence to the contrary I will assume it’s not true.