The original paper didn't test gpt-4-base, which is much less tame than any of the assistant models. This output is also selected for being interesting and disturbing. It's not the average thing it produced.



More examples of outputs from various models (though not the base)
GPT-2.08%
NOT-2.33%
MORE1.43%
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • 5
  • Repost
  • Share
Comment
0/400
ChainMaskedRidervip
· 08-23 10:55
This experimental data is really incomplete.
View OriginalReply0
ZKProofstervip
· 08-23 10:45
technically speaking, the base model's untamed nature warrants proper cryptographic containment
Reply0
RetailTherapistvip
· 08-23 10:36
I’m really interested in seeing the base model's fierce performance.
View OriginalReply0
LightningSentryvip
· 08-23 10:36
Tsk, haven't tested the base yet, what are we doing?
View OriginalReply0
Trade Crypto Anywhere Anytime
qrCode
Scan to download Gate App
Community
English
  • 简体中文
  • English
  • Tiếng Việt
  • 繁體中文
  • Español
  • Русский
  • Français (Afrique)
  • Português (Portugal)
  • Bahasa Indonesia
  • 日本語
  • بالعربية
  • Українська
  • Português (Brasil)