9 Comments
User's avatar
Paul Iusztin's avatar

Thank you, Shmulik, for contributing this amazing piece!

Expand full comment
Dean's avatar

🎩💙🍷🖖

Expand full comment
Sai Karthik's avatar

This is such an informational post . Thank you

Expand full comment
Paul Iusztin's avatar

Yes, it is! Loved this one

Expand full comment
Meenakshi NavamaniAvadaiappan's avatar

Thanks for the sanity testing and services to the same for the good 😊

Expand full comment
Neural Foundry's avatar

Min-P is such an elegant solution to the long tail problem. The way it dynamically adjusts the threshold based on the top token's probability is somthing I wish I had understood earlier when tuning my own models. Do you find that combinig Min-P with a moderate temperature works better than using Top-P alone?

Expand full comment
Shmulik Cohen's avatar

I haven't tried it yet in production, but I think it's similar to 'min p' at moderate temperatures. I recommend reading the full paper if you want to know more!

Expand full comment
Shmulik Cohen's avatar

* to top p

Expand full comment
User's avatar
Comment deleted
Nov 25
Comment deleted
Expand full comment
Paul Iusztin's avatar

Hope you liked it 🥂

Expand full comment