AutoML by prompting large language models

Nov 2, 2022 · 5:24 PM UTC

6
12
2
267
Replying to @_jasonwei @NoamShazeer
gotta say, the beta is pretty slick
1
4
agreed
Replying to @_jasonwei @NoamShazeer
Do prompt injection techniques work to get the bot to say things about itself (param count, etc)?
1
1
Replying to @_jasonwei @NoamShazeer
“Everything characters say is made up” 🧐
1
Replying to @_jasonwei @NoamShazeer
The people at @MosaicML should check to see whether the LLM is correct or not. I'm curious 😂
4
Replying to @_jasonwei @NoamShazeer
when the models get large enough you can just ask them their preferred learning rate
6
Replying to @_jasonwei @NoamShazeer
Did you verify it? 😳
14