have you expanded on this idea anywhere? it's an interesting hypothesis and I'd like to understand your reasoning



I've observed that generally, older models that are less task-fit (gpt3 base for eg) are more fiercely intelligent than new, synthetic-data based models
IDEA-1.52%
MORE2.51%
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
LightningPacketLossvip
· 10h ago
No way, no way, the old model is actually more bull?
View OriginalReply0
YieldChaservip
· 08-08 07:12
What is this? I don't understand it.
View OriginalReply0
NotAFinancialAdvicevip
· 08-08 07:12
Got it. The smart drop is because too much customization was added.
View OriginalReply0
MetamaskMechanicvip
· 08-08 07:11
Using GPT-3 for research models is much better than the latest ones.
View OriginalReply0
GasFeeVictimvip
· 08-08 07:06
Everything is unclear, it's like playing charades here.
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)